Techwizz/kernel-fxtec-pro1x
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
kernel-fxtec-pro1x/crypto/testmgr.h

32885 lines
1.1 MiB
C
Raw Normal View History

crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* Algorithm testing framework and tests.
*
* Copyright (c) 2002 James Morris <jmorris@intercode.com.au>
* Copyright (c) 2002 Jean-Francois Dive <jef@linuxbe.org>
* Copyright (c) 2007 Nokia Siemens Networks
* Copyright (c) 2008 Herbert Xu <herbert@gondor.apana.org.au>
*
crypto: rfc4106 - Extending the RC4106 AES-GCM test vectors Updated RFC4106 AES-GCM testing. Some test vectors were taken from http://csrc.nist.gov/groups/ST/toolkit/BCM/documents/proposedmodes/ gcm/gcm-test-vectors.tar.gz Signed-off-by: Adrian Hoban <adrian.hoban@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Gabriele Paoloni <gabriele.paoloni@intel.com> Signed-off-by: Aidan O'Mahony <aidan.o.mahony@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2010-11-04 13:02:04 -06:00
* Updated RFC4106 AES-GCM testing. Some test vectors were taken from
* http://csrc.nist.gov/groups/ST/toolkit/BCM/documents/proposedmodes/
* gcm/gcm-test-vectors.tar.gz
* Authors: Aidan O'Mahony (aidan.o.mahony@intel.com)
* Adrian Hoban <adrian.hoban@intel.com>
* Gabriele Paoloni <gabriele.paoloni@intel.com>
* Tadeusz Struk (tadeusz.struk@intel.com)
* Copyright (c) 2010, Intel Corporation.
*
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
* This program is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by the Free
* Software Foundation; either version 2 of the License, or (at your option)
* any later version.
*
*/
#ifndef _CRYPTO_TESTMGR_H
#define _CRYPTO_TESTMGR_H
crypto: testmgr - add zlib test Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-03-04 00:42:15 -07:00
#include <linux/netlink.h>
crypto: testmgr - Add support for the pcomp interface Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-03-04 00:08:03 -07:00
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
#define MAX_DIGEST_SIZE 64
#define MAX_TAP 8
crypto: testmgr - avoid DMA mapping from text, rodata, stack With DMA_API_DEBUG set, following warnings are emitted (tested on CAAM accelerator): DMA-API: device driver maps memory from kernel text or rodata DMA-API: device driver maps memory from stack and the culprits are: -key in __test_aead and __test_hash -result in __test_hash MAX_KEYLEN is changed to accommodate maximum key length from existing test vectors in crypto/testmgr.h (131 bytes) and rounded. Signed-off-by: Horia Geanta <horia.geanta@freescale.com> Acked-by: Kim Phillips <kim.phillips@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-23 02:59:38 -06:00
#define MAX_KEYLEN 160
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
#define MAX_IVLEN 32
struct hash_testvec {
/* only used with keyed hash algorithms */
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const char *key;
const char *plaintext;
const char *digest;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned char tap[MAX_TAP];
crypto: testmgr - add larger crc32c test vector to test FPU path in crc32c_intel Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-11-13 03:02:30 -07:00
unsigned short psize;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned char np;
unsigned char ksize;
};
crypto: testmgr - Document struct cipher_testvec Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-05 03:39:22 -06:00
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* cipher_testvec: structure to describe a symmetric cipher test
* @key: Pointer to key
* @klen: Length of @key in bytes
* @iv: Pointer to IV (optional for some ciphers)
* @ptext: Pointer to plaintext
* @ctext: Pointer to ciphertext
* @len: Length of @ptext and @ctext in bytes
crypto: testmgr - Document struct cipher_testvec Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-05 03:39:22 -06:00
* @fail: If set to one, the test need to fail
* @wk: Does the test need CRYPTO_TFM_REQ_WEAK_KEY
* ( e.g. test needs to fail due to a weak key )
* @np: numbers of SG to distribute data in (from 1 to MAX_TAP)
* @tap: How to distribute data in @np SGs
* @also_non_np: if set to 1, the test will be also done without
* splitting data in @np SGs
crypto: FIPS - allow tests to be disabled in FIPS mode In FIPS mode, additional restrictions may apply. If these restrictions are violated, the kernel will panic(). This patch allows test vectors for symmetric ciphers to be marked as to be skipped in FIPS mode. Together with the patch, the XTS test vectors where the AES key is identical to the tweak key is disabled in FIPS mode. This test vector violates the FIPS requirement that both keys must be different. Reported-by: Tapas Sarangi <TSarangi@trustwave.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-08-25 07:15:01 -06:00
* @fips_skip: Skip the test vector in FIPS mode
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* @generates_iv: Encryption should ignore the given IV, and output @iv.
* Decryption takes @iv. Needed for AES Keywrap ("kw(aes)").
crypto: testmgr - Document struct cipher_testvec Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-05 03:39:22 -06:00
*/
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
struct cipher_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const char *key;
const char *iv;
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
const char *ptext;
const char *ctext;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned short tap[MAX_TAP];
int np;
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
unsigned char also_non_np;
crypto: testmgr - Wrap the LHS in expressions of the form !x == y In the test manager, there are a number of if-statements with expressions of the form !x == y that incur warnings with gcc-5 of the following form: ../crypto/testmgr.c: In function '__test_aead': ../crypto/testmgr.c:523:12: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses] if (!ret == template[i].fail) { ^ By converting the 'fail' member of struct aead_testvec and struct cipher_testvec to a bool, we can get rid of the warnings. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-04-28 08:36:36 -06:00
bool fail;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned char wk; /* weak key flag */
unsigned char klen;
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
unsigned short len;
crypto: FIPS - allow tests to be disabled in FIPS mode In FIPS mode, additional restrictions may apply. If these restrictions are violated, the kernel will panic(). This patch allows test vectors for symmetric ciphers to be marked as to be skipped in FIPS mode. Together with the patch, the XTS test vectors where the AES key is identical to the tweak key is disabled in FIPS mode. This test vector violates the FIPS requirement that both keys must be different. Reported-by: Tapas Sarangi <TSarangi@trustwave.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-08-25 07:15:01 -06:00
bool fips_skip;
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
bool generates_iv;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
struct aead_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const char *key;
const char *iv;
const char *input;
const char *assoc;
const char *result;
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned char tap[MAX_TAP];
unsigned char atap[MAX_TAP];
int np;
int anp;
crypto: testmgr - Wrap the LHS in expressions of the form !x == y In the test manager, there are a number of if-statements with expressions of the form !x == y that incur warnings with gcc-5 of the following form: ../crypto/testmgr.c: In function '__test_aead': ../crypto/testmgr.c:523:12: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses] if (!ret == template[i].fail) { ^ By converting the 'fail' member of struct aead_testvec and struct cipher_testvec to a bool, we can get rid of the warnings. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-04-28 08:36:36 -06:00
bool fail;
crypto: testmgr - Handle AEAD test vectors expected to fail verification Add infrastructure to tcrypt/testmgr to support handling ccm decryption test vectors that are expected to fail verification. Signed-off-by: Jarod Wilson <jarod@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:22:11 -06:00
unsigned char novrfy; /* ccm dec verification failure expected */
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
unsigned char wk; /* weak key flag */
unsigned char klen;
unsigned short ilen;
unsigned short alen;
unsigned short rlen;
};
crypto: testmgr - Add infrastructure for ansi_cprng self-tests Add some necessary infrastructure to make it possible to run self-tests for ansi_cprng. The bits are likely very specific to the ANSI X9.31 CPRNG in AES mode, and thus perhaps should be named more specifically if/when we grow additional CPRNG support... Successfully tested against the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.x kernel with the follow-on patch that adds the actual test vectors. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:44:50 -06:00
struct cprng_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const char *key;
const char *dt;
const char *v;
const char *result;
crypto: testmgr - Add infrastructure for ansi_cprng self-tests Add some necessary infrastructure to make it possible to run self-tests for ansi_cprng. The bits are likely very specific to the ANSI X9.31 CPRNG in AES mode, and thus perhaps should be named more specifically if/when we grow additional CPRNG support... Successfully tested against the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.x kernel with the follow-on patch that adds the actual test vectors. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:44:50 -06:00
unsigned char klen;
unsigned short dtlen;
unsigned short vlen;
unsigned short rlen;
unsigned short loops;
};
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
struct drbg_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *entropy;
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
size_t entropylen;
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *entpra;
const unsigned char *entprb;
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
size_t entprlen;
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *addtla;
const unsigned char *addtlb;
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
size_t addtllen;
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *pers;
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
size_t perslen;
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *expected;
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
size_t expectedlen;
};
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
struct akcipher_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *key;
const unsigned char *m;
const unsigned char *c;
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
unsigned int key_len;
unsigned int m_size;
unsigned int c_size;
bool public_key_vec;
crypto: testmgr - add testvector for pkcs1pad(rsa) The PKCS#1 RSA implementation is provided with a self test with RSA 2048 and SHA-256. This self test implicitly covers other RSA keys and other hashes. Also, this self test implies that the pkcs1pad(rsa) is FIPS 140-2 compliant. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-06-12 15:27:51 -06:00
bool siggen_sigver_test;
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
};
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
struct kpp_testvec {
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *secret;
crypto: testmgr - add genkey kpp test The test considers a party that already has a private-public key pair and a party that provides a NULL key. The kernel will generate the private-public key pair for the latter, computes the shared secret on both ends and verifies if it's the same. The explicit private-public key pair was copied from the previous test vector. Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-05-30 08:52:49 -06:00
const unsigned char *b_secret;
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
const unsigned char *b_public;
const unsigned char *expected_a_public;
const unsigned char *expected_ss;
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
unsigned short secret_size;
crypto: testmgr - add genkey kpp test The test considers a party that already has a private-public key pair and a party that provides a NULL key. The kernel will generate the private-public key pair for the latter, computes the shared secret on both ends and verifies if it's the same. The explicit private-public key pair was copied from the previous test vector. Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-05-30 08:52:49 -06:00
unsigned short b_secret_size;
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
unsigned short b_public_size;
unsigned short expected_a_public_size;
unsigned short expected_ss_size;
crypto: testmgr - add genkey kpp test The test considers a party that already has a private-public key pair and a party that provides a NULL key. The kernel will generate the private-public key pair for the latter, computes the shared secret on both ends and verifies if it's the same. The explicit private-public key pair was copied from the previous test vector. Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-05-30 08:52:49 -06:00
bool genkey;
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const char zeroed_string[48];
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
/*
* RSA test vectors. Borrowed from openSSL.
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct akcipher_testvec rsa_tv_template[] = {
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
{
#ifndef CONFIG_CRYPTO_FIPS
.key =
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\x30\x81\x9A" /* sequence of 154 bytes */
"\x02\x01\x01" /* version - integer of 1 byte */
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
"\x02\x41" /* modulus - integer of 65 bytes */
"\x00\xAA\x36\xAB\xCE\x88\xAC\xFD\xFF\x55\x52\x3C\x7F\xC4\x52\x3F"
"\x90\xEF\xA0\x0D\xF3\x77\x4A\x25\x9F\x2E\x62\xB4\xC5\xD9\x9C\xB5"
"\xAD\xB3\x00\xA0\x28\x5E\x53\x01\x93\x0E\x0C\x70\xFB\x68\x76\x93"
"\x9C\xE6\x16\xCE\x62\x4A\x11\xE0\x08\x6D\x34\x1E\xBC\xAC\xA0\xA1"
"\xF5"
"\x02\x01\x11" /* public key - integer of 1 byte */
"\x02\x40" /* private key - integer of 64 bytes */
"\x0A\x03\x37\x48\x62\x64\x87\x69\x5F\x5F\x30\xBC\x38\xB9\x8B\x44"
"\xC2\xCD\x2D\xFF\x43\x40\x98\xCD\x20\xD8\xA1\x38\xD0\x90\xBF\x64"
"\x79\x7C\x3F\xA7\xA2\xCD\xCB\x3C\xD1\xE0\xBD\xBA\x26\x54\xB4\xF9"
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\xDF\x8E\x8A\xE5\x9D\x73\x3D\x9F\x33\xB3\x01\x62\x4A\xFD\x1D\x51"
"\x02\x01\x00" /* prime1 - integer of 1 byte */
"\x02\x01\x00" /* prime2 - integer of 1 byte */
"\x02\x01\x00" /* exponent1 - integer of 1 byte */
"\x02\x01\x00" /* exponent2 - integer of 1 byte */
"\x02\x01\x00", /* coefficient - integer of 1 byte */
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
.m = "\x54\x85\x9b\x34\x2c\x49\xea\x2a",
.c =
"\x63\x1c\xcd\x7b\xe1\x7e\xe4\xde\xc9\xa8\x89\xa1\x74\xcb\x3c\x63"
"\x7d\x24\xec\x83\xc3\x15\xe4\x7f\x73\x05\x34\xd1\xec\x22\xbb\x8a"
"\x5e\x32\x39\x6d\xc1\x1d\x7d\x50\x3b\x9f\x7a\xad\xf0\x2e\x25\x53"
"\x9f\x6e\xbd\x4c\x55\x84\x0c\x9b\xcf\x1a\x4b\x51\x1e\x9e\x0c\x06",
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
.key_len = 157,
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
.m_size = 8,
.c_size = 64,
}, {
.key =
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\x30\x82\x01\x1D" /* sequence of 285 bytes */
"\x02\x01\x01" /* version - integer of 1 byte */
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
"\x02\x81\x81" /* modulus - integer of 129 bytes */
"\x00\xBB\xF8\x2F\x09\x06\x82\xCE\x9C\x23\x38\xAC\x2B\x9D\xA8\x71"
"\xF7\x36\x8D\x07\xEE\xD4\x10\x43\xA4\x40\xD6\xB6\xF0\x74\x54\xF5"
"\x1F\xB8\xDF\xBA\xAF\x03\x5C\x02\xAB\x61\xEA\x48\xCE\xEB\x6F\xCD"
"\x48\x76\xED\x52\x0D\x60\xE1\xEC\x46\x19\x71\x9D\x8A\x5B\x8B\x80"
"\x7F\xAF\xB8\xE0\xA3\xDF\xC7\x37\x72\x3E\xE6\xB4\xB7\xD9\x3A\x25"
"\x84\xEE\x6A\x64\x9D\x06\x09\x53\x74\x88\x34\xB2\x45\x45\x98\x39"
"\x4E\xE0\xAA\xB1\x2D\x7B\x61\xA5\x1F\x52\x7A\x9A\x41\xF6\xC1\x68"
"\x7F\xE2\x53\x72\x98\xCA\x2A\x8F\x59\x46\xF8\xE5\xFD\x09\x1D\xBD"
"\xCB"
"\x02\x01\x11" /* public key - integer of 1 byte */
"\x02\x81\x81" /* private key - integer of 129 bytes */
"\x00\xA5\xDA\xFC\x53\x41\xFA\xF2\x89\xC4\xB9\x88\xDB\x30\xC1\xCD"
"\xF8\x3F\x31\x25\x1E\x06\x68\xB4\x27\x84\x81\x38\x01\x57\x96\x41"
"\xB2\x94\x10\xB3\xC7\x99\x8D\x6B\xC4\x65\x74\x5E\x5C\x39\x26\x69"
"\xD6\x87\x0D\xA2\xC0\x82\xA9\x39\xE3\x7F\xDC\xB8\x2E\xC9\x3E\xDA"
"\xC9\x7F\xF3\xAD\x59\x50\xAC\xCF\xBC\x11\x1C\x76\xF1\xA9\x52\x94"
"\x44\xE5\x6A\xAF\x68\xC5\x6C\x09\x2C\xD3\x8D\xC3\xBE\xF5\xD2\x0A"
"\x93\x99\x26\xED\x4F\x74\xA1\x3E\xDD\xFB\xE1\xA1\xCE\xCC\x48\x94"
"\xAF\x94\x28\xC2\xB7\xB8\x88\x3F\xE4\x46\x3A\x4B\xC8\x5B\x1C\xB3"
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\xC1"
"\x02\x01\x00" /* prime1 - integer of 1 byte */
"\x02\x01\x00" /* prime2 - integer of 1 byte */
"\x02\x01\x00" /* exponent1 - integer of 1 byte */
"\x02\x01\x00" /* exponent2 - integer of 1 byte */
"\x02\x01\x00", /* coefficient - integer of 1 byte */
.key_len = 289,
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
.m = "\x54\x85\x9b\x34\x2c\x49\xea\x2a",
.c =
"\x74\x1b\x55\xac\x47\xb5\x08\x0a\x6e\x2b\x2d\xf7\x94\xb8\x8a\x95"
"\xed\xa3\x6b\xc9\x29\xee\xb2\x2c\x80\xc3\x39\x3b\x8c\x62\x45\x72"
"\xc2\x7f\x74\x81\x91\x68\x44\x48\x5a\xdc\xa0\x7e\xa7\x0b\x05\x7f"
"\x0e\xa0\x6c\xe5\x8f\x19\x4d\xce\x98\x47\x5f\xbd\x5f\xfe\xe5\x34"
"\x59\x89\xaf\xf0\xba\x44\xd7\xf1\x1a\x50\x72\xef\x5e\x4a\xb6\xb7"
"\x54\x34\xd1\xc4\x83\x09\xdf\x0f\x91\x5f\x7d\x91\x70\x2f\xd4\x13"
"\xcc\x5e\xa4\x6c\xc3\x4d\x28\xef\xda\xaf\xec\x14\x92\xfc\xa3\x75"
"\x13\xb4\xc1\xa1\x11\xfc\x40\x2f\x4c\x9d\xdf\x16\x76\x11\x20\x6b",
.m_size = 8,
.c_size = 128,
}, {
#endif
.key =
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\x30\x82\x02\x1F" /* sequence of 543 bytes */
"\x02\x01\x01" /* version - integer of 1 byte */
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
"\x02\x82\x01\x00" /* modulus - integer of 256 bytes */
"\xDB\x10\x1A\xC2\xA3\xF1\xDC\xFF\x13\x6B\xED\x44\xDF\xF0\x02\x6D"
"\x13\xC7\x88\xDA\x70\x6B\x54\xF1\xE8\x27\xDC\xC3\x0F\x99\x6A\xFA"
"\xC6\x67\xFF\x1D\x1E\x3C\x1D\xC1\xB5\x5F\x6C\xC0\xB2\x07\x3A\x6D"
"\x41\xE4\x25\x99\xAC\xFC\xD2\x0F\x02\xD3\xD1\x54\x06\x1A\x51\x77"
"\xBD\xB6\xBF\xEA\xA7\x5C\x06\xA9\x5D\x69\x84\x45\xD7\xF5\x05\xBA"
"\x47\xF0\x1B\xD7\x2B\x24\xEC\xCB\x9B\x1B\x10\x8D\x81\xA0\xBE\xB1"
"\x8C\x33\xE4\x36\xB8\x43\xEB\x19\x2A\x81\x8D\xDE\x81\x0A\x99\x48"
"\xB6\xF6\xBC\xCD\x49\x34\x3A\x8F\x26\x94\xE3\x28\x82\x1A\x7C\x8F"
"\x59\x9F\x45\xE8\x5D\x1A\x45\x76\x04\x56\x05\xA1\xD0\x1B\x8C\x77"
"\x6D\xAF\x53\xFA\x71\xE2\x67\xE0\x9A\xFE\x03\xA9\x85\xD2\xC9\xAA"
"\xBA\x2A\xBC\xF4\xA0\x08\xF5\x13\x98\x13\x5D\xF0\xD9\x33\x34\x2A"
"\x61\xC3\x89\x55\xF0\xAE\x1A\x9C\x22\xEE\x19\x05\x8D\x32\xFE\xEC"
"\x9C\x84\xBA\xB7\xF9\x6C\x3A\x4F\x07\xFC\x45\xEB\x12\xE5\x7B\xFD"
"\x55\xE6\x29\x69\xD1\xC2\xE8\xB9\x78\x59\xF6\x79\x10\xC6\x4E\xEB"
"\x6A\x5E\xB9\x9A\xC7\xC4\x5B\x63\xDA\xA3\x3F\x5E\x92\x7A\x81\x5E"
"\xD6\xB0\xE2\x62\x8F\x74\x26\xC2\x0C\xD3\x9A\x17\x47\xE6\x8E\xAB"
"\x02\x03\x01\x00\x01" /* public key - integer of 3 bytes */
"\x02\x82\x01\x00" /* private key - integer of 256 bytes */
"\x52\x41\xF4\xDA\x7B\xB7\x59\x55\xCA\xD4\x2F\x0F\x3A\xCB\xA4\x0D"
"\x93\x6C\xCC\x9D\xC1\xB2\xFB\xFD\xAE\x40\x31\xAC\x69\x52\x21\x92"
"\xB3\x27\xDF\xEA\xEE\x2C\x82\xBB\xF7\x40\x32\xD5\x14\xC4\x94\x12"
"\xEC\xB8\x1F\xCA\x59\xE3\xC1\x78\xF3\x85\xD8\x47\xA5\xD7\x02\x1A"
"\x65\x79\x97\x0D\x24\xF4\xF0\x67\x6E\x75\x2D\xBF\x10\x3D\xA8\x7D"
"\xEF\x7F\x60\xE4\xE6\x05\x82\x89\x5D\xDF\xC6\xD2\x6C\x07\x91\x33"
"\x98\x42\xF0\x02\x00\x25\x38\xC5\x85\x69\x8A\x7D\x2F\x95\x6C\x43"
"\x9A\xB8\x81\xE2\xD0\x07\x35\xAA\x05\x41\xC9\x1E\xAF\xE4\x04\x3B"
"\x19\xB8\x73\xA2\xAC\x4B\x1E\x66\x48\xD8\x72\x1F\xAC\xF6\xCB\xBC"
"\x90\x09\xCA\xEC\x0C\xDC\xF9\x2C\xD7\xEB\xAE\xA3\xA4\x47\xD7\x33"
"\x2F\x8A\xCA\xBC\x5E\xF0\x77\xE4\x97\x98\x97\xC7\x10\x91\x7D\x2A"
"\xA6\xFF\x46\x83\x97\xDE\xE9\xE2\x17\x03\x06\x14\xE2\xD7\xB1\x1D"
"\x77\xAF\x51\x27\x5B\x5E\x69\xB8\x81\xE6\x11\xC5\x43\x23\x81\x04"
"\x62\xFF\xE9\x46\xB8\xD8\x44\xDB\xA5\xCC\x31\x54\x34\xCE\x3E\x82"
"\xD6\xBF\x7A\x0B\x64\x21\x6D\x88\x7E\x5B\x45\x12\x1E\x63\x8D\x49"
crypto: akcipher - Changes to asymmetric key API Setkey function has been split into set_priv_key and set_pub_key. Akcipher requests takes sgl for src and dst instead of void *. Users of the API i.e. two existing RSA implementation and test mgr code have been updated accordingly. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-10-08 10:26:55 -06:00
"\xA7\x1D\xD9\x1E\x06\xCD\xE8\xBA\x2C\x8C\x69\x32\xEA\xBE\x60\x71"
"\x02\x01\x00" /* prime1 - integer of 1 byte */
"\x02\x01\x00" /* prime2 - integer of 1 byte */
"\x02\x01\x00" /* exponent1 - integer of 1 byte */
"\x02\x01\x00" /* exponent2 - integer of 1 byte */
"\x02\x01\x00", /* coefficient - integer of 1 byte */
.key_len = 547,
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
.m = "\x54\x85\x9b\x34\x2c\x49\xea\x2a",
.c =
"\xb2\x97\x76\xb4\xae\x3e\x38\x3c\x7e\x64\x1f\xcc\xa2\x7f\xf6\xbe"
"\xcf\x49\xbc\x48\xd3\x6c\x8f\x0a\x0e\xc1\x73\xbd\x7b\x55\x79\x36"
"\x0e\xa1\x87\x88\xb9\x2c\x90\xa6\x53\x5e\xe9\xef\xc4\xe2\x4d\xdd"
"\xf7\xa6\x69\x82\x3f\x56\xa4\x7b\xfb\x62\xe0\xae\xb8\xd3\x04\xb3"
"\xac\x5a\x15\x2a\xe3\x19\x9b\x03\x9a\x0b\x41\xda\x64\xec\x0a\x69"
"\xfc\xf2\x10\x92\xf3\xc1\xbf\x84\x7f\xfd\x2c\xae\xc8\xb5\xf6\x41"
"\x70\xc5\x47\x03\x8a\xf8\xff\x6f\x3f\xd2\x6f\x09\xb4\x22\xf3\x30"
"\xbe\xa9\x85\xcb\x9c\x8d\xf9\x8f\xeb\x32\x91\xa2\x25\x84\x8f\xf5"
"\xdc\xc7\x06\x9c\x2d\xe5\x11\x2c\x09\x09\x87\x09\xa9\xf6\x33\x73"
"\x90\xf1\x60\xf2\x65\xdd\x30\xa5\x66\xce\x62\x7b\xd0\xf8\x2d\x3d"
"\x19\x82\x77\xe3\x0a\x5f\x75\x2f\x8e\xb1\xe5\xe8\x91\x35\x1b\x3b"
"\x33\xb7\x66\x92\xd1\xf2\x8e\x6f\xe5\x75\x0c\xad\x36\xfb\x4e\xd0"
"\x66\x61\xbd\x49\xfe\xf4\x1a\xa2\x2b\x49\xfe\x03\x4c\x74\x47\x8d"
"\x9a\x66\xb2\x49\x46\x4d\x77\xea\x33\x4d\x6b\x3c\xb4\x49\x4a\xc6"
"\x7d\x3d\xb5\xb9\x56\x41\x15\x67\x0f\x94\x3c\x93\x65\x27\xe0\x21"
"\x5d\x59\xc3\x62\xd5\xa6\xda\x38\x26\x22\x5e\x34\x1c\x94\xaf\x98",
.m_size = 8,
.c_size = 256,
}, {
.key =
"\x30\x82\x01\x09" /* sequence of 265 bytes */
"\x02\x82\x01\x00" /* modulus - integer of 256 bytes */
"\xDB\x10\x1A\xC2\xA3\xF1\xDC\xFF\x13\x6B\xED\x44\xDF\xF0\x02\x6D"
"\x13\xC7\x88\xDA\x70\x6B\x54\xF1\xE8\x27\xDC\xC3\x0F\x99\x6A\xFA"
"\xC6\x67\xFF\x1D\x1E\x3C\x1D\xC1\xB5\x5F\x6C\xC0\xB2\x07\x3A\x6D"
"\x41\xE4\x25\x99\xAC\xFC\xD2\x0F\x02\xD3\xD1\x54\x06\x1A\x51\x77"
"\xBD\xB6\xBF\xEA\xA7\x5C\x06\xA9\x5D\x69\x84\x45\xD7\xF5\x05\xBA"
"\x47\xF0\x1B\xD7\x2B\x24\xEC\xCB\x9B\x1B\x10\x8D\x81\xA0\xBE\xB1"
"\x8C\x33\xE4\x36\xB8\x43\xEB\x19\x2A\x81\x8D\xDE\x81\x0A\x99\x48"
"\xB6\xF6\xBC\xCD\x49\x34\x3A\x8F\x26\x94\xE3\x28\x82\x1A\x7C\x8F"
"\x59\x9F\x45\xE8\x5D\x1A\x45\x76\x04\x56\x05\xA1\xD0\x1B\x8C\x77"
"\x6D\xAF\x53\xFA\x71\xE2\x67\xE0\x9A\xFE\x03\xA9\x85\xD2\xC9\xAA"
"\xBA\x2A\xBC\xF4\xA0\x08\xF5\x13\x98\x13\x5D\xF0\xD9\x33\x34\x2A"
"\x61\xC3\x89\x55\xF0\xAE\x1A\x9C\x22\xEE\x19\x05\x8D\x32\xFE\xEC"
"\x9C\x84\xBA\xB7\xF9\x6C\x3A\x4F\x07\xFC\x45\xEB\x12\xE5\x7B\xFD"
"\x55\xE6\x29\x69\xD1\xC2\xE8\xB9\x78\x59\xF6\x79\x10\xC6\x4E\xEB"
"\x6A\x5E\xB9\x9A\xC7\xC4\x5B\x63\xDA\xA3\x3F\x5E\x92\x7A\x81\x5E"
"\xD6\xB0\xE2\x62\x8F\x74\x26\xC2\x0C\xD3\x9A\x17\x47\xE6\x8E\xAB"
"\x02\x03\x01\x00\x01", /* public key - integer of 3 bytes */
.key_len = 269,
.m = "\x54\x85\x9b\x34\x2c\x49\xea\x2a",
.c =
"\xb2\x97\x76\xb4\xae\x3e\x38\x3c\x7e\x64\x1f\xcc\xa2\x7f\xf6\xbe"
"\xcf\x49\xbc\x48\xd3\x6c\x8f\x0a\x0e\xc1\x73\xbd\x7b\x55\x79\x36"
"\x0e\xa1\x87\x88\xb9\x2c\x90\xa6\x53\x5e\xe9\xef\xc4\xe2\x4d\xdd"
"\xf7\xa6\x69\x82\x3f\x56\xa4\x7b\xfb\x62\xe0\xae\xb8\xd3\x04\xb3"
"\xac\x5a\x15\x2a\xe3\x19\x9b\x03\x9a\x0b\x41\xda\x64\xec\x0a\x69"
"\xfc\xf2\x10\x92\xf3\xc1\xbf\x84\x7f\xfd\x2c\xae\xc8\xb5\xf6\x41"
"\x70\xc5\x47\x03\x8a\xf8\xff\x6f\x3f\xd2\x6f\x09\xb4\x22\xf3\x30"
"\xbe\xa9\x85\xcb\x9c\x8d\xf9\x8f\xeb\x32\x91\xa2\x25\x84\x8f\xf5"
"\xdc\xc7\x06\x9c\x2d\xe5\x11\x2c\x09\x09\x87\x09\xa9\xf6\x33\x73"
"\x90\xf1\x60\xf2\x65\xdd\x30\xa5\x66\xce\x62\x7b\xd0\xf8\x2d\x3d"
"\x19\x82\x77\xe3\x0a\x5f\x75\x2f\x8e\xb1\xe5\xe8\x91\x35\x1b\x3b"
"\x33\xb7\x66\x92\xd1\xf2\x8e\x6f\xe5\x75\x0c\xad\x36\xfb\x4e\xd0"
"\x66\x61\xbd\x49\xfe\xf4\x1a\xa2\x2b\x49\xfe\x03\x4c\x74\x47\x8d"
"\x9a\x66\xb2\x49\x46\x4d\x77\xea\x33\x4d\x6b\x3c\xb4\x49\x4a\xc6"
"\x7d\x3d\xb5\xb9\x56\x41\x15\x67\x0f\x94\x3c\x93\x65\x27\xe0\x21"
"\x5d\x59\xc3\x62\xd5\xa6\xda\x38\x26\x22\x5e\x34\x1c\x94\xaf\x98",
.m_size = 8,
.c_size = 256,
.public_key_vec = true,
crypto: testmgr - use calculated count for number of test vectors When working on AES in CCM mode for ARM, my code passed the internal tcrypt test before I had even bothered to implement the AES-192 and AES-256 code paths, which is strange because the tcrypt does contain AES-192 and AES-256 test vectors for CCM. As it turned out, the define AES_CCM_ENC_TEST_VECTORS was out of sync with the actual number of test vectors, causing only the AES-128 ones to be executed. So get rid of the defines, and wrap the test vector references in a macro that calculates the number of vectors automatically. The following test vector counts were out of sync with the respective defines: BF_CTR_ENC_TEST_VECTORS 2 -> 3 BF_CTR_DEC_TEST_VECTORS 2 -> 3 TF_CTR_ENC_TEST_VECTORS 2 -> 3 TF_CTR_DEC_TEST_VECTORS 2 -> 3 SERPENT_CTR_ENC_TEST_VECTORS 2 -> 3 SERPENT_CTR_DEC_TEST_VECTORS 2 -> 3 AES_CCM_ENC_TEST_VECTORS 8 -> 14 AES_CCM_DEC_TEST_VECTORS 7 -> 17 AES_CCM_4309_ENC_TEST_VECTORS 7 -> 23 AES_CCM_4309_DEC_TEST_VECTORS 10 -> 23 CAMELLIA_CTR_ENC_TEST_VECTORS 2 -> 3 CAMELLIA_CTR_DEC_TEST_VECTORS 2 -> 3 Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-01-12 06:40:39 -07:00
#ifndef CONFIG_CRYPTO_FIPS
crypto: testmgr - Add 4K private key to RSA testvector Key generated with openssl. It also contains all fields required for testing CRT mode Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-07-04 10:21:39 -06:00
}, {
.key =
"\x30\x82\x09\x29" /* sequence of 2345 bytes */
"\x02\x01\x00" /* version integer of 1 byte */
"\x02\x82\x02\x01" /* modulus - integer of 513 bytes */
"\x00\xC3\x8B\x55\x7B\x73\x4D\xFF\xE9\x9B\xC6\xDC\x67\x3C\xB4\x8E"
"\xA0\x86\xED\xF2\xB9\x50\x5C\x54\x5C\xBA\xE4\xA1\xB2\xA7\xAE\x2F"
"\x1B\x7D\xF1\xFB\xAC\x79\xC5\xDF\x1A\x00\xC9\xB2\xC1\x61\x25\x33"
"\xE6\x9C\xE9\xCF\xD6\x27\xC4\x4E\x44\x30\x44\x5E\x08\xA1\x87\x52"
"\xCC\x6B\x97\x70\x8C\xBC\xA5\x06\x31\x0C\xD4\x2F\xD5\x7D\x26\x24"
"\xA2\xE2\xAC\x78\xF4\x53\x14\xCE\xF7\x19\x2E\xD7\xF7\xE6\x0C\xB9"
"\x56\x7F\x0B\xF1\xB1\xE2\x43\x70\xBD\x86\x1D\xA1\xCC\x2B\x19\x08"
"\x76\xEF\x91\xAC\xBF\x20\x24\x0D\x38\xC0\x89\xB8\x9A\x70\xB3\x64"
"\xD9\x8F\x80\x41\x10\x5B\x9F\xB1\xCB\x76\x43\x00\x21\x25\x36\xD4"
"\x19\xFC\x55\x95\x10\xE4\x26\x74\x98\x2C\xD9\xBD\x0B\x2B\x04\xC2"
"\xAC\x82\x38\xB4\xDD\x4C\x04\x7E\x51\x36\x40\x1E\x0B\xC4\x7C\x25"
"\xDD\x4B\xB2\xE7\x20\x0A\x57\xF9\xB4\x94\xC3\x08\x33\x22\x6F\x8B"
"\x48\xDB\x03\x68\x5A\x5B\xBA\xAE\xF3\xAD\xCF\xC3\x6D\xBA\xF1\x28"
"\x67\x7E\x6C\x79\x07\xDE\xFC\xED\xE7\x96\xE3\x6C\xE0\x2C\x87\xF8"
"\x02\x01\x28\x38\x43\x21\x53\x84\x69\x75\x78\x15\x7E\xEE\xD2\x1B"
"\xB9\x23\x40\xA8\x86\x1E\x38\x83\xB2\x73\x1D\x53\xFB\x9E\x2A\x8A"
"\xB2\x75\x35\x01\xC3\xC3\xC4\x94\xE8\x84\x86\x64\x81\xF4\x42\xAA"
"\x3C\x0E\xD6\x4F\xBC\x0A\x09\x2D\xE7\x1B\xD4\x10\xA8\x54\xEA\x89"
"\x84\x8A\xCB\xF7\x5A\x3C\xCA\x76\x08\x29\x62\xB4\x6A\x22\xDF\x14"
"\x95\x71\xFD\xB6\x86\x39\xB8\x8B\xF8\x91\x7F\x38\xAA\x14\xCD\xE5"
"\xF5\x1D\xC2\x6D\x53\x69\x52\x84\x7F\xA3\x1A\x5E\x26\x04\x83\x06"
"\x73\x52\x56\xCF\x76\x26\xC9\xDD\x75\xD7\xFC\xF4\x69\xD8\x7B\x55"
"\xB7\x68\x13\x53\xB9\xE7\x89\xC3\xE8\xD6\x6E\xA7\x6D\xEA\x81\xFD"
"\xC4\xB7\x05\x5A\xB7\x41\x0A\x23\x8E\x03\x8A\x1C\xAE\xD3\x1E\xCE"
"\xE3\x5E\xFC\x19\x4A\xEE\x61\x9B\x8E\xE5\xE5\xDD\x85\xF9\x41\xEC"
"\x14\x53\x92\xF7\xDD\x06\x85\x02\x91\xE3\xEB\x6C\x43\x03\xB1\x36"
"\x7B\x89\x5A\xA8\xEB\xFC\xD5\xA8\x35\xDC\x81\xD9\x5C\xBD\xCA\xDC"
"\x9B\x98\x0B\x06\x5D\x0C\x5B\xEE\xF3\xD5\xCC\x57\xC9\x71\x2F\x90"
"\x3B\x3C\xF0\x8E\x4E\x35\x48\xAE\x63\x74\xA9\xFC\x72\x75\x8E\x34"
"\xA8\xF2\x1F\xEA\xDF\x3A\x37\x2D\xE5\x39\x39\xF8\x57\x58\x3C\x04"
"\xFE\x87\x06\x98\xBC\x7B\xD3\x21\x36\x60\x25\x54\xA7\x3D\xFA\x91"
"\xCC\xA8\x0B\x92\x8E\xB4\xF7\x06\xFF\x1E\x95\xCB\x07\x76\x97\x3B"
"\x9D"
"\x02\x03\x01\x00\x01" /* public key integer of 3 bytes */
"\x02\x82\x02\x00" /* private key integer of 512 bytes */
"\x74\xA9\xE0\x6A\x32\xB4\xCA\x85\xD9\x86\x9F\x60\x88\x7B\x40\xCC"
"\xCD\x33\x91\xA8\xB6\x25\x1F\xBF\xE3\x51\x1C\x97\xB6\x2A\xD9\xB8"
"\x11\x40\x19\xE3\x21\x13\xC8\xB3\x7E\xDC\xD7\x65\x40\x4C\x2D\xD6"
"\xDC\xAF\x32\x6C\x96\x75\x2C\x2C\xCA\x8F\x3F\x7A\xEE\xC4\x09\xC6"
"\x24\x3A\xC9\xCF\x6D\x8D\x17\x50\x94\x52\xD3\xE7\x0F\x2F\x7E\x94"
"\x1F\xA0\xBE\xD9\x25\xE8\x38\x42\x7C\x27\xD2\x79\xF8\x2A\x87\x38"
"\xEF\xBB\x74\x8B\xA8\x6E\x8C\x08\xC6\xC7\x4F\x0C\xBC\x79\xC6\xEF"
"\x0E\xA7\x5E\xE4\xF8\x8C\x09\xC7\x5E\x37\xCC\x87\x77\xCD\xCF\xD1"
"\x6D\x28\x1B\xA9\x62\xC0\xB8\x16\xA7\x8B\xF9\xBB\xCC\xB4\x15\x7F"
"\x1B\x69\x03\xF2\x7B\xEB\xE5\x8C\x14\xD6\x23\x4F\x52\x6F\x18\xA6"
"\x4B\x5B\x01\xAD\x35\xF9\x48\x53\xB3\x86\x35\x66\xD7\xE7\x29\xC0"
"\x09\xB5\xC6\xE6\xFA\xC4\xDA\x19\xBE\xD7\x4D\x41\x14\xBE\x6F\xDF"
"\x1B\xAB\xC0\xCA\x88\x07\xAC\xF1\x7D\x35\x83\x67\x28\x2D\x50\xE9"
"\xCE\x27\x71\x5E\x1C\xCF\xD2\x30\x65\x79\x72\x2F\x9C\xE1\xD2\x39"
"\x7F\xEF\x3B\x01\xF2\x14\x1D\xDF\xBD\x51\xD3\xA1\x53\x62\xCF\x5F"
"\x79\x84\xCE\x06\x96\x69\x29\x49\x82\x1C\x71\x4A\xA1\x66\xC8\x2F"
"\xFD\x7B\x96\x7B\xFC\xC4\x26\x58\xC4\xFC\x7C\xAF\xB5\xE8\x95\x83"
"\x87\xCB\x46\xDE\x97\xA7\xB3\xA2\x54\x5B\xD7\xAF\xAB\xEB\xC8\xF3"
"\x55\x9D\x48\x2B\x30\x9C\xDC\x26\x4B\xC2\x89\x45\x13\xB2\x01\x9A"
"\xA4\x65\xC3\xEC\x24\x2D\x26\x97\xEB\x80\x8A\x9D\x03\xBC\x59\x66"
"\x9E\xE2\xBB\xBB\x63\x19\x64\x93\x11\x7B\x25\x65\x30\xCD\x5B\x4B"
"\x2C\xFF\xDC\x2D\x30\x87\x1F\x3C\x88\x07\xD0\xFC\x48\xCC\x05\x8A"
"\xA2\xC8\x39\x3E\xD5\x51\xBC\x0A\xBE\x6D\xA8\xA0\xF6\x88\x06\x79"
"\x13\xFF\x1B\x45\xDA\x54\xC9\x24\x25\x8A\x75\x0A\x26\xD1\x69\x81"
"\x14\x14\xD1\x79\x7D\x8E\x76\xF2\xE0\xEB\xDD\x0F\xDE\xC2\xEC\x80"
"\xD7\xDC\x16\x99\x92\xBE\xCB\x40\x0C\xCE\x7C\x3B\x46\xA2\x5B\x5D"
"\x0C\x45\xEB\xE1\x00\xDE\x72\x50\xB1\xA6\x0B\x76\xC5\x8D\xFC\x82"
"\x38\x6D\x99\x14\x1D\x1A\x4A\xD3\x7C\x53\xB8\x12\x46\xA2\x30\x38"
"\x82\xF4\x96\x6E\x8C\xCE\x47\x0D\xAF\x0A\x3B\x45\xB7\x43\x95\x43"
"\x9E\x02\x2C\x44\x07\x6D\x1F\x3C\x66\x89\x09\xB6\x1F\x06\x30\xCC"
"\xAD\xCE\x7D\x9A\xDE\x3E\xFB\x6C\xE4\x58\x43\xD2\x4F\xA5\x9E\x5E"
"\xA7\x7B\xAE\x3A\xF6\x7E\xD9\xDB\xD3\xF5\xC5\x41\xAF\xE6\x9C\x91"
"\x02\x82\x01\x01" /* prime1 - integer of 257 bytes */
"\x00\xE0\xA6\x6C\xF0\xA2\xF8\x81\x85\x36\x43\xD0\x13\x0B\x33\x8B"
"\x8F\x78\x3D\xAC\xC7\x5E\x46\x6A\x7F\x05\xAE\x3E\x26\x0A\xA6\xD0"
"\x51\xF3\xC8\x61\xF5\x77\x22\x48\x10\x87\x4C\xD5\xA4\xD5\xAE\x2D"
"\x4E\x7A\xFE\x1C\x31\xE7\x6B\xFF\xA4\x69\x20\xF9\x2A\x0B\x99\xBE"
"\x7C\x32\x68\xAD\xB0\xC6\x94\x81\x41\x75\xDC\x06\x78\x0A\xB4\xCF"
"\xCD\x1B\x2D\x31\xE4\x7B\xEA\xA8\x35\x99\x75\x57\xC6\x0E\xF6\x78"
"\x4F\xA0\x92\x4A\x00\x1B\xE7\x96\xF2\x5B\xFD\x2C\x0A\x0A\x13\x81"
"\xAF\xCB\x59\x87\x31\xD9\x83\x65\xF2\x22\x48\xD0\x03\x67\x39\xF6"
"\xFF\xA8\x36\x07\x3A\x68\xE3\x7B\xA9\x64\xFD\x9C\xF7\xB1\x3D\xBF"
"\x26\x5C\xCC\x7A\xFC\xA2\x8F\x51\xD1\xE1\xE2\x3C\xEC\x06\x75\x7C"
"\x34\xF9\xA9\x33\x70\x11\xAD\x5A\xDC\x5F\xCF\x50\xF6\x23\x2F\x39"
"\xAC\x92\x48\x53\x4D\x01\x96\x3C\xD8\xDC\x1F\x23\x23\x78\x80\x34"
"\x54\x14\x76\x8B\xB6\xBB\xFB\x88\x78\x31\x59\x28\xD2\xB1\x75\x17"
"\x88\x04\x4A\x78\x62\x18\x2E\xF5\xFB\x9B\xEF\x15\xD8\x16\x47\xC6"
"\x42\xB1\x02\xDA\x9E\xE3\x84\x90\xB4\x2D\xC3\xCE\x13\xC9\x12\x7D"
"\x3E\xCD\x39\x39\xC9\xAD\xA1\x1A\xE6\xD5\xAD\x5A\x09\x4D\x1B\x0C"
"\xAB"
"\x02\x82\x01\x01" /* prime 2 - integer of 257 bytes */
"\x00\xDE\xD5\x1B\xF6\xCD\x83\xB1\xC6\x47\x7E\xB9\xC0\x6B\xA9\xB8"
"\x02\xF3\xAE\x40\x5D\xFC\xD3\xE5\x4E\xF1\xE3\x39\x04\x52\x84\x89"
"\x40\x37\xBB\xC2\xCD\x7F\x71\x77\x17\xDF\x6A\x4C\x31\x24\x7F\xB9"
"\x7E\x7F\xC8\x43\x4A\x3C\xEB\x8D\x1B\x7F\x21\x51\x67\x45\x8F\xA0"
"\x36\x29\x3A\x18\x45\xA5\x32\xEC\x74\x88\x3C\x98\x5D\x67\x3B\xD7"
"\x51\x1F\xE9\xAE\x09\x01\xDE\xDE\x7C\xFB\x60\xD1\xA5\x6C\xE9\x6A"
"\x93\x04\x02\x3A\xBB\x67\x02\xB9\xFD\x23\xF0\x02\x2B\x49\x85\xC9"
"\x5B\xE7\x4B\xDF\xA3\xF4\xEE\x59\x4C\x45\xEF\x8B\xC1\x6B\xDE\xDE"
"\xBC\x1A\xFC\xD2\x76\x3F\x33\x74\xA9\x8E\xA3\x7E\x0C\xC6\xCE\x70"
"\xA1\x5B\xA6\x77\xEA\x76\xEB\x18\xCE\xB9\xD7\x78\x8D\xAE\x06\xBB"
"\xD3\x1F\x16\x0D\x05\xAB\x4F\xC6\x52\xC8\x6B\x36\x51\x7D\x1D\x27"
"\xAF\x88\x9A\x6F\xCC\x25\x2E\x74\x06\x72\xCE\x9E\xDB\xE0\x9D\x30"
"\xEF\x55\xA5\x58\x21\xA7\x42\x12\x2C\x2C\x23\x87\xC1\x0F\xE8\x51"
"\xDA\x53\xDA\xFC\x05\x36\xDF\x08\x0E\x08\x36\xBE\x5C\x86\x9E\xCA"
"\x68\x90\x33\x12\x0B\x14\x82\xAB\x90\x1A\xD4\x49\x32\x9C\xBD\xAA"
"\xAB\x4E\x38\xF1\xEE\xED\x3D\x3F\xE8\xBD\x48\x56\xA6\x64\xEE\xC8"
"\xD7"
"\x02\x82\x01\x01" /* exponent 1 - integer of 257 bytes */
"\x00\x96\x5E\x6F\x8F\x06\xD6\xE6\x03\x1F\x96\x76\x81\x38\xBF\x30"
"\xCC\x40\x84\xAF\xD0\xE7\x06\xA5\x24\x0E\xCE\x59\xA5\x26\xFE\x0F"
"\x74\xBB\x83\xC6\x26\x02\xAF\x3C\xA3\x6B\x9C\xFF\x68\x0C\xEB\x40"
"\x42\x46\xCB\x2E\x5E\x2C\xF4\x3A\x32\x77\x77\xED\xAF\xBA\x02\x17"
"\xE1\x93\xF0\x43\x4A\x8F\x31\x39\xEF\x72\x0F\x6B\x79\x10\x59\x84"
"\xBA\x5A\x55\x7F\x0E\xDB\xEE\xEE\xD6\xA9\xB8\x44\x9F\x3A\xC6\xB9"
"\x33\x3B\x5C\x90\x11\xD0\x9B\xCC\x8A\xBF\x0E\x10\x5B\x4B\xF1\x50"
"\x9E\x35\xB3\xE0\x6D\x7A\x95\x9C\x38\x5D\xC0\x75\x13\xC2\x15\xA7"
"\x81\xEA\xBA\xF7\x4D\x9E\x85\x9D\xF1\x7D\xBA\xD0\x45\x6F\x2A\xD0"
"\x76\xC2\x28\xD0\xAD\xA7\xB5\xDC\xE3\x6A\x99\xFF\x83\x50\xB3\x75"
"\x07\x14\x91\xAF\xEF\x74\xB5\x9F\x9A\xE0\xBA\xA9\x0B\x87\xF3\x85"
"\x5C\x40\xB2\x0E\xA7\xFD\xC6\xED\x45\x8E\xD9\x7C\xB0\xB2\x68\xC6"
"\x1D\xFD\x70\x78\x06\x41\x7F\x95\x12\x36\x9D\xE2\x58\x5D\x15\xEE"
"\x41\x49\xF5\xFA\xEC\x56\x19\xA0\xE6\xE0\xB2\x40\xE1\xD9\xD0\x03"
"\x22\x02\xCF\xD1\x3C\x07\x38\x65\x8F\x65\x0E\xAA\x32\xCE\x25\x05"
"\x16\x73\x51\xB9\x9F\x88\x0B\xCD\x30\xF3\x97\xCC\x2B\x6B\xA4\x0E"
"\x6F"
"\x02\x82\x01\x00" /* exponent 2 - integer of 256 bytes */
"\x2A\x5F\x3F\xB8\x08\x90\x58\x47\xA9\xE4\xB1\x11\xA3\xE7\x5B\xF4"
"\x43\xBE\x08\xC3\x56\x86\x3C\x7E\x6C\x84\x96\x9C\xF9\xCB\xF6\x05"
"\x5E\x13\xB8\x11\x37\x80\xAD\xF2\xBE\x2B\x0A\x5D\xF5\xE0\xCB\xB7"
"\x00\x39\x66\x82\x41\x5F\x51\x2F\xBF\x56\xE8\x91\xC8\xAA\x6C\xFE"
"\x9F\x8C\x4A\x7D\x43\xD2\x91\x1F\xFF\x9F\xF6\x21\x1C\xB6\x46\x55"
"\x48\xCA\x38\xAB\xC1\xCD\x4D\x65\x5A\xAF\xA8\x6D\xDA\x6D\xF0\x34"
"\x10\x79\x14\x0D\xFA\xA2\x8C\x17\x54\xB4\x18\xD5\x7E\x5F\x90\x50"
"\x87\x84\xE7\xFB\xD7\x61\x53\x5D\xAB\x96\xC7\x6E\x7A\x42\xA0\xFC"
"\x07\xED\xB7\x5F\x80\xD9\x19\xFF\xFB\xFD\x9E\xC4\x73\x31\x62\x3D"
"\x6C\x9E\x15\x03\x62\xA5\x85\xCC\x19\x8E\x9D\x7F\xE3\x6D\xA8\x5D"
"\x96\xF5\xAC\x78\x3D\x81\x27\xE7\x29\xF1\x29\x1D\x09\xBB\x77\x86"
"\x6B\x65\x62\x88\xE1\x31\x1A\x22\xF7\xC5\xCE\x73\x65\x1C\xBE\xE7"
"\x63\xD3\xD3\x14\x63\x27\xAF\x28\xF3\x23\xB6\x76\xC1\xBD\x9D\x82"
"\xF4\x9B\x19\x7D\x2C\x57\xF0\xC2\x2A\x51\xAE\x95\x0D\x8C\x38\x54"
"\xF5\xC6\xA0\x51\xB7\x0E\xB9\xEC\xE7\x0D\x22\xF6\x1A\xD3\xFE\x16"
"\x21\x03\xB7\x0D\x85\xD3\x35\xC9\xDD\xE4\x59\x85\xBE\x7F\xA1\x75"
"\x02\x82\x01\x01" /* coefficient - integer of 257 bytes */
"\x00\xB9\x48\xD2\x54\x2F\x19\x54\x64\xAE\x62\x80\x61\x89\x80\xB4"
"\x48\x0B\x8D\x7E\x1B\x0F\x50\x08\x82\x3F\xED\x75\x84\xB7\x13\xE4"
"\xF8\x8D\xA8\xBB\x54\x21\x4C\x5A\x54\x07\x16\x4B\xB4\xA4\x9E\x30"
"\xBF\x7A\x30\x1B\x39\x60\xA3\x21\x53\xFB\xB0\xDC\x0F\x7C\x2C\xFB"
"\xAA\x95\x7D\x51\x39\x28\x33\x1F\x25\x31\x53\xF5\xD2\x64\x2B\xF2"
"\x1E\xB3\xC0\x6A\x0B\xC9\xA4\x42\x64\x5C\xFB\x15\xA3\xE8\x4C\x3A"
"\x9C\x3C\xBE\xA3\x39\x83\x23\xE3\x6D\x18\xCC\xC2\xDC\x63\x8D\xBA"
"\x98\xE0\xE0\x31\x4A\x2B\x37\x9C\x4D\x6B\xF3\x9F\x51\xE4\x43\x5C"
"\x83\x5F\xBF\x5C\xFE\x92\x45\x01\xAF\xF5\xC2\xF4\xB7\x56\x93\xA5"
"\xF4\xAA\x67\x3C\x48\x37\xBD\x9A\x3C\xFE\xA5\x9A\xB0\xD1\x6B\x85"
"\xDD\x81\xD4\xFA\xAD\x31\x83\xA8\x22\x9B\xFD\xB4\x61\xDC\x7A\x51"
"\x59\x62\x10\x1B\x7E\x44\xA3\xFE\x90\x51\x5A\x3E\x02\x87\xAD\xFA"
"\xDD\x0B\x1F\x3D\x35\xAF\xEE\x13\x85\x51\xA7\x42\xC0\xEE\x9E\x20"
"\xE9\xD0\x29\xB2\xE4\x21\xE4\x6D\x62\xB9\xF4\x48\x4A\xD8\x46\x8E"
"\x61\xA6\x2C\x5D\xDF\x8F\x97\x2B\x3A\x75\x1D\x83\x17\x6F\xC6\xB0"
"\xDE\xFC\x14\x25\x06\x5A\x60\xBB\xB8\x21\x89\xD1\xEF\x57\xF1\x71"
"\x3D",
.m = "\x54\x85\x9b\x34\x2c\x49\xea\x2a",
.c =
"\x5c\xce\x9c\xd7\x9a\x9e\xa1\xfe\x7a\x82\x3c\x68\x27\x98\xe3\x5d"
"\xd5\xd7\x07\x29\xf5\xfb\xc3\x1a\x7f\x63\x1e\x62\x31\x3b\x19\x87"
"\x79\x4f\xec\x7b\xf3\xcb\xea\x9b\x95\x52\x3a\x40\xe5\x87\x7b\x72"
"\xd1\x72\xc9\xfb\x54\x63\xd8\xc9\xd7\x2c\xfc\x7b\xc3\x14\x1e\xbc"
"\x18\xb4\x34\xa1\xbf\x14\xb1\x37\x31\x6e\xf0\x1b\x35\x19\x54\x07"
"\xf7\x99\xec\x3e\x63\xe2\xcd\x61\x28\x65\xc3\xcd\xb1\x38\x36\xa5"
"\xb2\xd7\xb0\xdc\x1f\xf5\xef\x19\xc7\x53\x32\x2d\x1c\x26\xda\xe4"
"\x0d\xd6\x90\x7e\x28\xd8\xdc\xe4\x61\x05\xd2\x25\x90\x01\xd3\x96"
"\x6d\xa6\xcf\x58\x20\xbb\x03\xf4\x01\xbc\x79\xb9\x18\xd8\xb8\xba"
"\xbd\x93\xfc\xf2\x62\x5d\x8c\x66\x1e\x0e\x84\x59\x93\xdd\xe2\x93"
"\xa2\x62\x7d\x08\x82\x7a\xdd\xfc\xb8\xbc\xc5\x4f\x9c\x4e\xbf\xb4"
"\xfc\xf4\xc5\x01\xe8\x00\x70\x4d\x28\x26\xcc\x2e\xfe\x0e\x58\x41"
"\x8b\xec\xaf\x7c\x4b\x54\xd0\xa0\x64\xf9\x32\xf4\x2e\x47\x65\x0a"
"\x67\x88\x39\x3a\xdb\xb2\xdb\x7b\xb5\xf6\x17\xa8\xd9\xc6\x5e\x28"
"\x13\x82\x8a\x99\xdb\x60\x08\xa5\x23\x37\xfa\x88\x90\x31\xc8\x9d"
"\x8f\xec\xfb\x85\x9f\xb1\xce\xa6\x24\x50\x46\x44\x47\xcb\x65\xd1"
"\xdf\xc0\xb1\x6c\x90\x1f\x99\x8e\x4d\xd5\x9e\x31\x07\x66\x87\xdf"
"\x01\xaa\x56\x3c\x71\xe0\x2b\x6f\x67\x3b\x23\xed\xc2\xbd\x03\x30"
"\x79\x76\x02\x10\x10\x98\x85\x8a\xff\xfd\x0b\xda\xa5\xd9\x32\x48"
"\x02\xa0\x0b\xb9\x2a\x8a\x18\xca\xc6\x8f\x3f\xbb\x16\xb2\xaa\x98"
"\x27\xe3\x60\x43\xed\x15\x70\xd4\x57\x15\xfe\x19\xd4\x9b\x13\x78"
"\x8a\xf7\x21\xf1\xa2\xa2\x2d\xb3\x09\xcf\x44\x91\x6e\x08\x3a\x30"
"\x81\x3e\x90\x93\x8a\x67\x33\x00\x59\x54\x9a\x25\xd3\x49\x8e\x9f"
"\xc1\x4b\xe5\x86\xf3\x50\x4c\xbc\xc5\xd3\xf5\x3a\x54\xe1\x36\x3f"
"\xe2\x5a\xb4\x37\xc0\xeb\x70\x35\xec\xf6\xb7\xe8\x44\x3b\x7b\xf3"
"\xf1\xf2\x1e\xdb\x60\x7d\xd5\xbe\xf0\x71\x34\x90\x4c\xcb\xd4\x35"
"\x51\xc7\xdd\xd8\xc9\x81\xf5\x5d\x57\x46\x2c\xb1\x7b\x9b\xaa\xcb"
"\xd1\x22\x25\x49\x44\xa3\xd4\x6b\x29\x7b\xd8\xb2\x07\x93\xbf\x3d"
"\x52\x49\x84\x79\xef\xb8\xe5\xc4\xad\xca\xa8\xc6\xf6\xa6\x76\x70"
"\x5b\x0b\xe5\x83\xc6\x0e\xef\x55\xf2\xe7\xff\x04\xea\xe6\x13\xbe"
"\x40\xe1\x40\x45\x48\x66\x75\x31\xae\x35\x64\x91\x11\x6f\xda\xee"
"\x26\x86\x45\x6f\x0b\xd5\x9f\x03\xb1\x65\x5b\xdb\xa4\xe4\xf9\x45",
.key_len = 2349,
.m_size = 8,
.c_size = 512,
crypto: testmgr - use calculated count for number of test vectors When working on AES in CCM mode for ARM, my code passed the internal tcrypt test before I had even bothered to implement the AES-192 and AES-256 code paths, which is strange because the tcrypt does contain AES-192 and AES-256 test vectors for CCM. As it turned out, the define AES_CCM_ENC_TEST_VECTORS was out of sync with the actual number of test vectors, causing only the AES-128 ones to be executed. So get rid of the defines, and wrap the test vector references in a macro that calculates the number of vectors automatically. The following test vector counts were out of sync with the respective defines: BF_CTR_ENC_TEST_VECTORS 2 -> 3 BF_CTR_DEC_TEST_VECTORS 2 -> 3 TF_CTR_ENC_TEST_VECTORS 2 -> 3 TF_CTR_DEC_TEST_VECTORS 2 -> 3 SERPENT_CTR_ENC_TEST_VECTORS 2 -> 3 SERPENT_CTR_DEC_TEST_VECTORS 2 -> 3 AES_CCM_ENC_TEST_VECTORS 8 -> 14 AES_CCM_DEC_TEST_VECTORS 7 -> 17 AES_CCM_4309_ENC_TEST_VECTORS 7 -> 23 AES_CCM_4309_DEC_TEST_VECTORS 10 -> 23 CAMELLIA_CTR_ENC_TEST_VECTORS 2 -> 3 CAMELLIA_CTR_DEC_TEST_VECTORS 2 -> 3 Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-01-12 06:40:39 -07:00
#endif
crypto: testmgr - add tests vectors for RSA New test vectors for RSA algorithm. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 11:31:06 -06:00
}
};
crypto: testmgr - add testvector for pkcs1pad(rsa) The PKCS#1 RSA implementation is provided with a self test with RSA 2048 and SHA-256. This self test implicitly covers other RSA keys and other hashes. Also, this self test implies that the pkcs1pad(rsa) is FIPS 140-2 compliant. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-06-12 15:27:51 -06:00
/*
* PKCS#1 RSA test vectors. Obtained from CAVS testing.
*/
static const struct akcipher_testvec pkcs1pad_rsa_tv_template[] = {
{
.key =
crypto: testmgr - Fix incorrect values in PKCS#1 test vector The RSA private key for the first form should have version, prime1, prime2, exponent1, exponent2, coefficient values 0. With non-zero values for prime1,2, exponent 1,2 and coefficient the Intel QAT driver will assume that values are provided for the private key second form. This will result in signature verification failures for modules where QAT device is present and the modules are signed with rsa,sha256. Cc: <stable@vger.kernel.org> Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Conor McLoughlin <conor.mcloughlin@intel.com> Reviewed-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-13 01:29:56 -07:00
"\x30\x82\x03\x1f\x02\x01\x00\x02\x82\x01\x01\x00\xd7\x1e\x77\x82"
crypto: testmgr - add testvector for pkcs1pad(rsa) The PKCS#1 RSA implementation is provided with a self test with RSA 2048 and SHA-256. This self test implicitly covers other RSA keys and other hashes. Also, this self test implies that the pkcs1pad(rsa) is FIPS 140-2 compliant. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-06-12 15:27:51 -06:00
"\x8c\x92\x31\xe7\x69\x02\xa2\xd5\x5c\x78\xde\xa2\x0c\x8f\xfe\x28"
"\x59\x31\xdf\x40\x9c\x60\x61\x06\xb9\x2f\x62\x40\x80\x76\xcb\x67"
"\x4a\xb5\x59\x56\x69\x17\x07\xfa\xf9\x4c\xbd\x6c\x37\x7a\x46\x7d"
"\x70\xa7\x67\x22\xb3\x4d\x7a\x94\xc3\xba\x4b\x7c\x4b\xa9\x32\x7c"
"\xb7\x38\x95\x45\x64\xa4\x05\xa8\x9f\x12\x7c\x4e\xc6\xc8\x2d\x40"
"\x06\x30\xf4\x60\xa6\x91\xbb\x9b\xca\x04\x79\x11\x13\x75\xf0\xae"
"\xd3\x51\x89\xc5\x74\xb9\xaa\x3f\xb6\x83\xe4\x78\x6b\xcd\xf9\x5c"
"\x4c\x85\xea\x52\x3b\x51\x93\xfc\x14\x6b\x33\x5d\x30\x70\xfa\x50"
"\x1b\x1b\x38\x81\x13\x8d\xf7\xa5\x0c\xc0\x8e\xf9\x63\x52\x18\x4e"
"\xa9\xf9\xf8\x5c\x5d\xcd\x7a\x0d\xd4\x8e\x7b\xee\x91\x7b\xad\x7d"
"\xb4\x92\xd5\xab\x16\x3b\x0a\x8a\xce\x8e\xde\x47\x1a\x17\x01\x86"
"\x7b\xab\x99\xf1\x4b\x0c\x3a\x0d\x82\x47\xc1\x91\x8c\xbb\x2e\x22"
"\x9e\x49\x63\x6e\x02\xc1\xc9\x3a\x9b\xa5\x22\x1b\x07\x95\xd6\x10"
"\x02\x50\xfd\xfd\xd1\x9b\xbe\xab\xc2\xc0\x74\xd7\xec\x00\xfb\x11"
"\x71\xcb\x7a\xdc\x81\x79\x9f\x86\x68\x46\x63\x82\x4d\xb7\xf1\xe6"
"\x16\x6f\x42\x63\xf4\x94\xa0\xca\x33\xcc\x75\x13\x02\x82\x01\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01\x00\x01"
"\x02\x82\x01\x00\x62\xb5\x60\x31\x4f\x3f\x66\x16\xc1\x60\xac\x47"
"\x2a\xff\x6b\x69\x00\x4a\xb2\x5c\xe1\x50\xb9\x18\x74\xa8\xe4\xdc"
"\xa8\xec\xcd\x30\xbb\xc1\xc6\xe3\xc6\xac\x20\x2a\x3e\x5e\x8b\x12"
"\xe6\x82\x08\x09\x38\x0b\xab\x7c\xb3\xcc\x9c\xce\x97\x67\xdd\xef"
"\x95\x40\x4e\x92\xe2\x44\xe9\x1d\xc1\x14\xfd\xa9\xb1\xdc\x71\x9c"
"\x46\x21\xbd\x58\x88\x6e\x22\x15\x56\xc1\xef\xe0\xc9\x8d\xe5\x80"
"\x3e\xda\x7e\x93\x0f\x52\xf6\xf5\xc1\x91\x90\x9e\x42\x49\x4f\x8d"
"\x9c\xba\x38\x83\xe9\x33\xc2\x50\x4f\xec\xc2\xf0\xa8\xb7\x6e\x28"
"\x25\x56\x6b\x62\x67\xfe\x08\xf1\x56\xe5\x6f\x0e\x99\xf1\xe5\x95"
"\x7b\xef\xeb\x0a\x2c\x92\x97\x57\x23\x33\x36\x07\xdd\xfb\xae\xf1"
"\xb1\xd8\x33\xb7\x96\x71\x42\x36\xc5\xa4\xa9\x19\x4b\x1b\x52\x4c"
"\x50\x69\x91\xf0\x0e\xfa\x80\x37\x4b\xb5\xd0\x2f\xb7\x44\x0d\xd4"
"\xf8\x39\x8d\xab\x71\x67\x59\x05\x88\x3d\xeb\x48\x48\x33\x88\x4e"
"\xfe\xf8\x27\x1b\xd6\x55\x60\x5e\x48\xb7\x6d\x9a\xa8\x37\xf9\x7a"
"\xde\x1b\xcd\x5d\x1a\x30\xd4\xe9\x9e\x5b\x3c\x15\xf8\x9c\x1f\xda"
"\xd1\x86\x48\x55\xce\x83\xee\x8e\x51\xc7\xde\x32\x12\x47\x7d\x46"
crypto: testmgr - Fix incorrect values in PKCS#1 test vector The RSA private key for the first form should have version, prime1, prime2, exponent1, exponent2, coefficient values 0. With non-zero values for prime1,2, exponent 1,2 and coefficient the Intel QAT driver will assume that values are provided for the private key second form. This will result in signature verification failures for modules where QAT device is present and the modules are signed with rsa,sha256. Cc: <stable@vger.kernel.org> Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Conor McLoughlin <conor.mcloughlin@intel.com> Reviewed-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-13 01:29:56 -07:00
"\xb8\x35\xdf\x41\x02\x01\x00\x02\x01\x00\x02\x01\x00\x02\x01\x00"
"\x02\x01\x00",
crypto: testmgr - add testvector for pkcs1pad(rsa) The PKCS#1 RSA implementation is provided with a self test with RSA 2048 and SHA-256. This self test implicitly covers other RSA keys and other hashes. Also, this self test implies that the pkcs1pad(rsa) is FIPS 140-2 compliant. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-06-12 15:27:51 -06:00
.key_len = 804,
/*
* m is SHA256 hash of following message:
* "\x49\x41\xbe\x0a\x0c\xc9\xf6\x35\x51\xe4\x27\x56\x13\x71\x4b\xd0"
* "\x36\x92\x84\x89\x1b\xf8\x56\x4a\x72\x61\x14\x69\x4f\x5e\x98\xa5"
* "\x80\x5a\x37\x51\x1f\xd8\xf5\xb5\x63\xfc\xf4\xb1\xbb\x4d\x33\xa3"
* "\x1e\xb9\x75\x8b\x9c\xda\x7e\x6d\x3a\x77\x85\xf7\xfc\x4e\xe7\x64"
* "\x43\x10\x19\xa0\x59\xae\xe0\xad\x4b\xd3\xc4\x45\xf7\xb1\xc2\xc1"
* "\x65\x01\x41\x39\x5b\x45\x47\xed\x2b\x51\xed\xe3\xd0\x09\x10\xd2"
* "\x39\x6c\x4a\x3f\xe5\xd2\x20\xe6\xb0\x71\x7d\x5b\xed\x26\x60\xf1"
* "\xb4\x73\xd1\xdb\x7d\xc4\x19\x91\xee\xf6\x32\x76\xf2\x19\x7d\xb7"
*/
.m =
"\x3e\xc8\xa1\x26\x20\x54\x44\x52\x48\x0d\xe5\x66\xf3\xb3\xf5\x04"
"\xbe\x10\xa8\x48\x94\x22\x2d\xdd\xba\x7a\xb4\x76\x8d\x79\x98\x89",
.m_size = 32,
.c =
"\xc7\xa3\x98\xeb\x43\xd1\x08\xc2\x3d\x78\x45\x04\x70\xc9\x01\xee"
"\xf8\x85\x37\x7c\x0b\xf9\x19\x70\x5c\x45\x7b\x2f\x3a\x0b\xb7\x8b"
"\xc4\x0d\x7b\x3a\x64\x0b\x0f\xdb\x78\xa9\x0b\xfd\x8d\x82\xa4\x86"
"\x39\xbf\x21\xb8\x84\xc4\xce\x9f\xc2\xe8\xb6\x61\x46\x17\xb9\x4e"
"\x0b\x57\x05\xb4\x4f\xf9\x9c\x93\x2d\x9b\xd5\x48\x1d\x80\x12\xef"
"\x3a\x77\x7f\xbc\xb5\x8e\x2b\x6b\x7c\xfc\x9f\x8c\x9d\xa2\xc4\x85"
"\xb0\x87\xe9\x17\x9b\xb6\x23\x62\xd2\xa9\x9f\x57\xe8\xf7\x04\x45"
"\x24\x3a\x45\xeb\xeb\x6a\x08\x8e\xaf\xc8\xa0\x84\xbc\x5d\x13\x38"
"\xf5\x17\x8c\xa3\x96\x9b\xa9\x38\x8d\xf0\x35\xad\x32\x8a\x72\x5b"
"\xdf\x21\xab\x4b\x0e\xa8\x29\xbb\x61\x54\xbf\x05\xdb\x84\x84\xde"
"\xdd\x16\x36\x31\xda\xf3\x42\x6d\x7a\x90\x22\x9b\x11\x29\xa6\xf8"
"\x30\x61\xda\xd3\x8b\x54\x1e\x42\xd1\x47\x1d\x6f\xd1\xcd\x42\x0b"
"\xd1\xe4\x15\x85\x7e\x08\xd6\x59\x64\x4c\x01\x34\x91\x92\x26\xe8"
"\xb0\x25\x8c\xf8\xf4\xfa\x8b\xc9\x31\x33\x76\x72\xfb\x64\x92\x9f"
"\xda\x62\x8d\xe1\x2a\x71\x91\x43\x40\x61\x3c\x5a\xbe\x86\xfc\x5b"
"\xe6\xf9\xa9\x16\x31\x1f\xaf\x25\x6d\xc2\x4a\x23\x6e\x63\x02\xa2",
.c_size = 256,
.siggen_sigver_test = true,
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct kpp_testvec dh_tv_template[] = {
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
{
.secret =
#ifdef __LITTLE_ENDIAN
"\x01\x00" /* type */
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
"\x15\x02" /* len */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x01\x00\x00" /* key_size */
"\x00\x01\x00\x00" /* p_size */
crypto: dh - update test for public key verification By adding a zero byte-length for the DH parameter Q value, the public key verification test is disabled for the given test. Reported-by: Eric Biggers <ebiggers3@gmail.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-11 12:35:49 -06:00
"\x00\x00\x00\x00" /* q_size */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x01\x00\x00\x00" /* g_size */
#else
"\x00\x01" /* type */
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
"\x02\x15" /* len */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x00\x01\x00" /* key_size */
"\x00\x00\x01\x00" /* p_size */
crypto: dh - update test for public key verification By adding a zero byte-length for the DH parameter Q value, the public key verification test is disabled for the given test. Reported-by: Eric Biggers <ebiggers3@gmail.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-11 12:35:49 -06:00
"\x00\x00\x00\x00" /* q_size */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x00\x00\x01" /* g_size */
#endif
/* xa */
"\x44\xc1\x48\x36\xa7\x2b\x6f\x4e\x43\x03\x68\xad\x31\x00\xda\xf3"
"\x2a\x01\xa8\x32\x63\x5f\x89\x32\x1f\xdf\x4c\xa1\x6a\xbc\x10\x15"
"\x90\x35\xc9\x26\x41\xdf\x7b\xaa\x56\x56\x3d\x85\x44\xb5\xc0\x8e"
"\x37\x83\x06\x50\xb3\x5f\x0e\x28\x2c\xd5\x46\x15\xe3\xda\x7d\x74"
"\x87\x13\x91\x4f\xd4\x2d\xf6\xc7\x5e\x14\x2c\x11\xc2\x26\xb4\x3a"
"\xe3\xb2\x36\x20\x11\x3b\x22\xf2\x06\x65\x66\xe2\x57\x58\xf8\x22"
"\x1a\x94\xbd\x2b\x0e\x8c\x55\xad\x61\x23\x45\x2b\x19\x1e\x63\x3a"
"\x13\x61\xe3\xa0\x79\x70\x3e\x6d\x98\x32\xbc\x7f\x82\xc3\x11\xd8"
"\xeb\x53\xb5\xfc\xb5\xd5\x3c\x4a\xea\x92\x3e\x01\xce\x15\x65\xd4"
"\xaa\x85\xc1\x11\x90\x83\x31\x6e\xfe\xe7\x7f\x7d\xed\xab\xf9\x29"
"\xf8\xc7\xf1\x68\xc6\xb7\xe4\x1f\x2f\x28\xa0\xc9\x1a\x50\x64\x29"
"\x4b\x01\x6d\x1a\xda\x46\x63\x21\x07\x40\x8c\x8e\x4c\x6f\xb5\xe5"
"\x12\xf3\xc2\x1b\x48\x27\x5e\x27\x01\xb1\xaa\xed\x68\x9b\x83\x18"
"\x8f\xb1\xeb\x1f\x04\xd1\x3c\x79\xed\x4b\xf7\x0a\x33\xdc\xe0\xc6"
"\xd8\x02\x51\x59\x00\x74\x30\x07\x4c\x2d\xac\xe4\x13\xf1\x80\xf0"
"\xce\xfa\xff\xa9\xce\x29\x46\xdd\x9d\xad\xd1\xc3\xc6\x58\x1a\x63"
/* p */
"\xb9\x36\x3a\xf1\x82\x1f\x60\xd3\x22\x47\xb8\xbc\x2d\x22\x6b\x81"
"\x7f\xe8\x20\x06\x09\x23\x73\x49\x9a\x59\x8b\x35\x25\xf8\x31\xbc"
"\x7d\xa8\x1c\x9d\x56\x0d\x1a\xf7\x4b\x4f\x96\xa4\x35\x77\x6a\x89"
"\xab\x42\x00\x49\x21\x71\xed\x28\x16\x1d\x87\x5a\x10\xa7\x9c\x64"
"\x94\xd4\x87\x3d\x28\xef\x44\xfe\x4b\xe2\xb4\x15\x8c\x82\xa6\xf3"
"\x50\x5f\xa8\xe8\xa2\x60\xe7\x00\x86\x78\x05\xd4\x78\x19\xa1\x98"
"\x62\x4e\x4a\x00\x78\x56\x96\xe6\xcf\xd7\x10\x1b\x74\x5d\xd0\x26"
"\x61\xdb\x6b\x32\x09\x51\xd8\xa5\xfd\x54\x16\x71\x01\xb3\x39\xe6"
"\x4e\x69\xb1\xd7\x06\x8f\xd6\x1e\xdc\x72\x25\x26\x74\xc8\x41\x06"
"\x5c\xd1\x26\x5c\xb0\x2f\xf9\x59\x13\xc1\x2a\x0f\x78\xea\x7b\xf7"
"\xbd\x59\xa0\x90\x1d\xfc\x33\x5b\x4c\xbf\x05\x9c\x3a\x3f\x69\xa2"
"\x45\x61\x4e\x10\x6a\xb3\x17\xc5\x68\x30\xfb\x07\x5f\x34\xc6\xfb"
"\x73\x07\x3c\x70\xf6\xae\xe7\x72\x84\xc3\x18\x81\x8f\xe8\x11\x1f"
"\x3d\x83\x83\x01\x2a\x14\x73\xbf\x32\x32\x2e\xc9\x4d\xdb\x2a\xca"
"\xee\x71\xf9\xda\xad\xe8\x82\x0b\x4d\x0c\x1f\xb6\x1d\xef\x00\x67"
"\x74\x3d\x95\xe0\xb7\xc4\x30\x8a\x24\x87\x12\x47\x27\x70\x0d\x73"
/* g */
"\x02",
.b_public =
"\x2a\x67\x5c\xfd\x63\x5d\xc0\x97\x0a\x8b\xa2\x1f\xf8\x8a\xcb\x54"
"\xca\x2f\xd3\x49\x3f\x01\x8e\x87\xfe\xcc\x94\xa0\x3e\xd4\x26\x79"
"\x9a\x94\x3c\x11\x81\x58\x5c\x60\x3d\xf5\x98\x90\x89\x64\x62\x1f"
"\xbd\x05\x6d\x2b\xcd\x84\x40\x9b\x4a\x1f\xe0\x19\xf1\xca\x20\xb3"
"\x4e\xa0\x4f\x15\xcc\xa5\xfe\xa5\xb4\xf5\x0b\x18\x7a\x5a\x37\xaa"
"\x58\x00\x19\x7f\xe2\xa3\xd9\x1c\x44\x57\xcc\xde\x2e\xc1\x38\xea"
"\xeb\xe3\x90\x40\xc4\x6c\xf7\xcd\xe9\x22\x50\x71\xf5\x7c\xdb\x37"
"\x0e\x80\xc3\xed\x7e\xb1\x2b\x2f\xbe\x71\xa6\x11\xa5\x9d\xf5\x39"
"\xf1\xa2\xe5\x85\xbc\x25\x91\x4e\x84\x8d\x26\x9f\x4f\xe6\x0f\xa6"
"\x2b\x6b\xf9\x0d\xaf\x6f\xbb\xfa\x2d\x79\x15\x31\x57\xae\x19\x60"
"\x22\x0a\xf5\xfd\x98\x0e\xbf\x5d\x49\x75\x58\x37\xbc\x7f\xf5\x21"
"\x56\x1e\xd5\xb3\x50\x0b\xca\x96\xf3\xd1\x3f\xb3\x70\xa8\x6d\x63"
"\x48\xfb\x3d\xd7\x29\x91\x45\xb5\x48\xcd\xb6\x78\x30\xf2\x3f\x1e"
"\xd6\x22\xd6\x35\x9b\xf9\x1f\x85\xae\xab\x4b\xd7\xe0\xc7\x86\x67"
"\x3f\x05\x7f\xa6\x0d\x2f\x0d\xbf\x53\x5f\x4d\x2c\x6d\x5e\x57\x40"
"\x30\x3a\x23\x98\xf9\xb4\x32\xf5\x32\x83\xdd\x0b\xae\x33\x97\x2f",
.expected_a_public =
"\x5c\x24\xdf\xeb\x5b\x4b\xf8\xc5\xef\x39\x48\x82\xe0\x1e\x62\xee"
"\x8a\xae\xdf\x93\x6c\x2b\x16\x95\x92\x16\x3f\x16\x7b\x75\x03\x85"
"\xd9\xf1\x69\xc2\x14\x87\x45\xfc\xa4\x19\xf6\xf0\xa4\xf3\xec\xd4"
"\x6c\x5c\x03\x3b\x94\xc2\x2f\x92\xe4\xce\xb3\xe4\x72\xe8\x17\xe6"
"\x23\x7e\x00\x01\x09\x59\x13\xbf\xc1\x2f\x99\xa9\x07\xaa\x02\x23"
"\x4a\xca\x39\x4f\xbc\xec\x0f\x27\x4f\x19\x93\x6c\xb9\x30\x52\xfd"
"\x2b\x9d\x86\xf1\x06\x1e\xb6\x56\x27\x4a\xc9\x8a\xa7\x8a\x48\x5e"
"\xb5\x60\xcb\xdf\xff\x03\x26\x10\xbf\x90\x8f\x46\x60\xeb\x9b\x9a"
"\xd6\x6f\x44\x91\x03\x92\x18\x2c\x96\x5e\x40\x19\xfb\xf4\x4f\x3a"
"\x02\x7b\xaf\xcc\x22\x20\x79\xb9\xf8\x9f\x8f\x85\x6b\xec\x44\xbb"
"\xe6\xa8\x8e\xb1\xe8\x2c\xee\x64\xee\xf8\xbd\x00\xf3\xe2\x2b\x93"
"\xcd\xe7\xc4\xdf\xc9\x19\x46\xfe\xb6\x07\x73\xc1\x8a\x64\x79\x26"
"\xe7\x30\xad\x2a\xdf\xe6\x8f\x59\xf5\x81\xbf\x4a\x29\x91\xe7\xb7"
"\xcf\x48\x13\x27\x75\x79\x40\xd9\xd6\x32\x52\x4e\x6a\x86\xae\x6f"
"\xc2\xbf\xec\x1f\xc2\x69\xb2\xb6\x59\xe5\xa5\x17\xa4\x77\xb7\x62"
"\x46\xde\xe8\xd2\x89\x78\x9a\xef\xa3\xb5\x8f\x26\xec\x80\xda\x39",
.expected_ss =
"\x8f\xf3\xac\xa2\xea\x22\x11\x5c\x45\x65\x1a\x77\x75\x2e\xcf\x46"
"\x23\x14\x1e\x67\x53\x4d\x35\xb0\x38\x1d\x4e\xb9\x41\x9a\x21\x24"
"\x6e\x9f\x40\xfe\x90\x51\xb1\x06\xa4\x7b\x87\x17\x2f\xe7\x5e\x22"
"\xf0\x7b\x54\x84\x0a\xac\x0a\x90\xd2\xd7\xe8\x7f\xe7\xe3\x30\x75"
"\x01\x1f\x24\x75\x56\xbe\xcc\x8d\x1e\x68\x0c\x41\x72\xd3\xfa\xbb"
"\xe5\x9c\x60\xc7\x28\x77\x0c\xbe\x89\xab\x08\xd6\x21\xe7\x2e\x1a"
"\x58\x7a\xca\x4f\x22\xf3\x2b\x30\xfd\xf4\x98\xc1\xa3\xf8\xf6\xcc"
"\xa9\xe4\xdb\x5b\xee\xd5\x5c\x6f\x62\x4c\xd1\x1a\x02\x2a\x23\xe4"
"\xb5\x57\xf3\xf9\xec\x04\x83\x54\xfe\x08\x5e\x35\xac\xfb\xa8\x09"
"\x82\x32\x60\x11\xb2\x16\x62\x6b\xdf\xda\xde\x9c\xcb\x63\x44\x6c"
"\x59\x26\x6a\x8f\xb0\x24\xcb\xa6\x72\x48\x1e\xeb\xe0\xe1\x09\x44"
"\xdd\xee\x66\x6d\x84\xcf\xa5\xc1\xb8\x36\x74\xd3\x15\x96\xc3\xe4"
"\xc6\x5a\x4d\x23\x97\x0c\x5c\xcb\xa9\xf5\x29\xc2\x0e\xff\x93\x82"
"\xd3\x34\x49\xad\x64\xa6\xb1\xc0\x59\x28\x75\x60\xa7\x8a\xb0\x11"
"\x56\x89\x42\x74\x11\xf5\xf6\x5e\x6f\x16\x54\x6a\xb1\x76\x4d\x50"
"\x8a\x68\xc1\x5b\x82\xb9\x0d\x00\x32\x50\xed\x88\x87\x48\x92\x17",
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
.secret_size = 533,
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
.b_public_size = 256,
.expected_a_public_size = 256,
.expected_ss_size = 256,
},
{
.secret =
#ifdef __LITTLE_ENDIAN
"\x01\x00" /* type */
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
"\x15\x02" /* len */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x01\x00\x00" /* key_size */
"\x00\x01\x00\x00" /* p_size */
crypto: dh - update test for public key verification By adding a zero byte-length for the DH parameter Q value, the public key verification test is disabled for the given test. Reported-by: Eric Biggers <ebiggers3@gmail.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-11 12:35:49 -06:00
"\x00\x00\x00\x00" /* q_size */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x01\x00\x00\x00" /* g_size */
#else
"\x00\x01" /* type */
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
"\x02\x15" /* len */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x00\x01\x00" /* key_size */
"\x00\x00\x01\x00" /* p_size */
crypto: dh - update test for public key verification By adding a zero byte-length for the DH parameter Q value, the public key verification test is disabled for the given test. Reported-by: Eric Biggers <ebiggers3@gmail.com> Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-11 12:35:49 -06:00
"\x00\x00\x00\x00" /* q_size */
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
"\x00\x00\x00\x01" /* g_size */
#endif
/* xa */
"\x4d\x75\xa8\x6e\xba\x23\x3a\x0c\x63\x56\xc8\xc9\x5a\xa7\xd6\x0e"
"\xed\xae\x40\x78\x87\x47\x5f\xe0\xa7\x7b\xba\x84\x88\x67\x4e\xe5"
"\x3c\xcc\x5c\x6a\xe7\x4a\x20\xec\xbe\xcb\xf5\x52\x62\x9f\x37\x80"
"\x0c\x72\x7b\x83\x66\xa4\xf6\x7f\x95\x97\x1c\x6a\x5c\x7e\xf1\x67"
"\x37\xb3\x93\x39\x3d\x0b\x55\x35\xd9\xe5\x22\x04\x9f\xf8\xc1\x04"
"\xce\x13\xa5\xac\xe1\x75\x05\xd1\x2b\x53\xa2\x84\xef\xb1\x18\xf4"
"\x66\xdd\xea\xe6\x24\x69\x5a\x49\xe0\x7a\xd8\xdf\x1b\xb7\xf1\x6d"
"\x9b\x50\x2c\xc8\x1c\x1c\xa3\xb4\x37\xfb\x66\x3f\x67\x71\x73\xa9"
"\xff\x5f\xd9\xa2\x25\x6e\x25\x1b\x26\x54\xbf\x0c\xc6\xdb\xea\x0a"
"\x52\x6c\x16\x7c\x27\x68\x15\x71\x58\x73\x9d\xe6\xc2\x80\xaa\x97"
"\x31\x66\xfb\xa6\xfb\xfd\xd0\x9c\x1d\xbe\x81\x48\xf5\x9a\x32\xf1"
"\x69\x62\x18\x78\xae\x72\x36\xe6\x94\x27\xd1\xff\x18\x4f\x28\x6a"
"\x16\xbd\x6a\x60\xee\xe5\xf9\x6d\x16\xe4\xb8\xa6\x41\x9b\x23\x7e"
"\xf7\x9d\xd1\x1d\x03\x15\x66\x3a\xcf\xb6\x2c\x13\x96\x2c\x52\x21"
"\xe4\x2d\x48\x7a\x8a\x5d\xb2\x88\xed\x98\x61\x79\x8b\x6a\x1e\x5f"
"\xd0\x8a\x2d\x99\x5a\x2b\x0f\xbc\xef\x53\x8f\x32\xc1\xa2\x99\x26"
/* p */
"\xb9\x36\x3a\xf1\x82\x1f\x60\xd3\x22\x47\xb8\xbc\x2d\x22\x6b\x81"
"\x7f\xe8\x20\x06\x09\x23\x73\x49\x9a\x59\x8b\x35\x25\xf8\x31\xbc"
"\x7d\xa8\x1c\x9d\x56\x0d\x1a\xf7\x4b\x4f\x96\xa4\x35\x77\x6a\x89"
"\xab\x42\x00\x49\x21\x71\xed\x28\x16\x1d\x87\x5a\x10\xa7\x9c\x64"
"\x94\xd4\x87\x3d\x28\xef\x44\xfe\x4b\xe2\xb4\x15\x8c\x82\xa6\xf3"
"\x50\x5f\xa8\xe8\xa2\x60\xe7\x00\x86\x78\x05\xd4\x78\x19\xa1\x98"
"\x62\x4e\x4a\x00\x78\x56\x96\xe6\xcf\xd7\x10\x1b\x74\x5d\xd0\x26"
"\x61\xdb\x6b\x32\x09\x51\xd8\xa5\xfd\x54\x16\x71\x01\xb3\x39\xe6"
"\x4e\x69\xb1\xd7\x06\x8f\xd6\x1e\xdc\x72\x25\x26\x74\xc8\x41\x06"
"\x5c\xd1\x26\x5c\xb0\x2f\xf9\x59\x13\xc1\x2a\x0f\x78\xea\x7b\xf7"
"\xbd\x59\xa0\x90\x1d\xfc\x33\x5b\x4c\xbf\x05\x9c\x3a\x3f\x69\xa2"
"\x45\x61\x4e\x10\x6a\xb3\x17\xc5\x68\x30\xfb\x07\x5f\x34\xc6\xfb"
"\x73\x07\x3c\x70\xf6\xae\xe7\x72\x84\xc3\x18\x81\x8f\xe8\x11\x1f"
"\x3d\x83\x83\x01\x2a\x14\x73\xbf\x32\x32\x2e\xc9\x4d\xdb\x2a\xca"
"\xee\x71\xf9\xda\xad\xe8\x82\x0b\x4d\x0c\x1f\xb6\x1d\xef\x00\x67"
"\x74\x3d\x95\xe0\xb7\xc4\x30\x8a\x24\x87\x12\x47\x27\x70\x0d\x73"
/* g */
"\x02",
.b_public =
"\x99\x4d\xd9\x01\x84\x8e\x4a\x5b\xb8\xa5\x64\x8c\x6c\x00\x5c\x0e"
"\x1e\x1b\xee\x5d\x9f\x53\xe3\x16\x70\x01\xed\xbf\x4f\x14\x36\x6e"
"\xe4\x43\x45\x43\x49\xcc\xb1\xb0\x2a\xc0\x6f\x22\x55\x42\x17\x94"
"\x18\x83\xd7\x2a\x5c\x51\x54\xf8\x4e\x7c\x10\xda\x76\x68\x57\x77"
"\x1e\x62\x03\x30\x04\x7b\x4c\x39\x9c\x54\x01\x54\xec\xef\xb3\x55"
"\xa4\xc0\x24\x6d\x3d\xbd\xcc\x46\x5b\x00\x96\xc7\xea\x93\xd1\x3f"
"\xf2\x6a\x72\xe3\xf2\xc1\x92\x24\x5b\xda\x48\x70\x2c\xa9\x59\x97"
"\x19\xb1\xd6\x54\xb3\x9c\x2e\xb0\x63\x07\x9b\x5e\xac\xb5\xf2\xb1"
"\x5b\xf8\xf3\xd7\x2d\x37\x9b\x68\x6c\xf8\x90\x07\xbc\x37\x9a\xa5"
"\xe2\x91\x12\x25\x47\x77\xe3\x3d\xb2\x95\x69\x44\x0b\x91\x1e\xaf"
"\x7c\x8c\x7c\x34\x41\x6a\xab\x60\x6e\xc6\x52\xec\x7e\x94\x0a\x37"
"\xec\x98\x90\xdf\x3f\x02\xbd\x23\x52\xdd\xd9\xe5\x31\x80\x74\x25"
"\xb6\xd2\xd3\xcc\xd5\xcc\x6d\xf9\x7e\x4d\x78\xab\x77\x51\xfa\x77"
"\x19\x94\x49\x8c\x05\xd4\x75\xed\xd2\xb3\x64\x57\xe0\x52\x99\xc0"
"\x83\xe3\xbb\x5e\x2b\xf1\xd2\xc0\xb1\x37\x36\x0b\x7c\xb5\x63\x96"
"\x8e\xde\x04\x23\x11\x95\x62\x11\x9a\xce\x6f\x63\xc8\xd5\xd1\x8f",
.expected_a_public =
"\x90\x89\xe4\x82\xd6\x0a\xcf\x1a\xae\xce\x1b\x66\xa7\x19\x71\x18"
"\x8f\x95\x4b\x5b\x80\x45\x4a\x5a\x43\x99\x4d\x37\xcf\xa3\xa7\x28"
"\x9c\xc7\x73\xf1\xb2\x17\xf6\x99\xe3\x6b\x56\xcb\x3e\x35\x60\x7d"
"\x65\xc7\x84\x6b\x3e\x60\xee\xcd\xd2\x70\xe7\xc9\x32\x1c\xf0\xb4"
"\xf9\x52\xd9\x88\x75\xfd\x40\x2c\xa7\xbe\x19\x1c\x0a\xae\x93\xe1"
"\x71\xc7\xcd\x4f\x33\x5c\x10\x7d\x39\x56\xfc\x73\x84\xb2\x67\xc3"
"\x77\x26\x20\x97\x2b\xf8\x13\x43\x93\x9c\x9a\xa4\x08\xc7\x34\x83"
"\xe6\x98\x61\xe7\x16\x30\x2c\xb1\xdb\x2a\xb2\xcc\xc3\x02\xa5\x3c"
"\x71\x50\x14\x83\xc7\xbb\xa4\xbe\x98\x1b\xfe\xcb\x43\xe9\x97\x62"
"\xd6\xf0\x8c\xcb\x1c\xba\x1e\xa8\xa6\xa6\x50\xfc\x85\x7d\x47\xbf"
"\xf4\x3e\x23\xd3\x5f\xb2\x71\x3e\x40\x94\xaa\x87\x83\x2c\x6c\x8e"
"\x60\xfd\xdd\xf7\xf4\x76\x03\xd3\x1d\xec\x18\x51\xa3\xf2\x44\x1a"
"\x3f\xb4\x7c\x18\x0d\x68\x65\x92\x54\x0d\x2d\x81\x16\xf1\x84\x66"
"\x89\x92\xd0\x1a\x5e\x1f\x42\x46\x5b\xe5\x83\x86\x80\xd9\xcd\x3a"
"\x5a\x2f\xb9\x59\x9b\xe4\x43\x84\x64\xf3\x09\x1a\x0a\xa2\x64\x0f"
"\x77\x4e\x8d\x8b\xe6\x88\xd1\xfc\xaf\x8f\xdf\x1d\xbc\x31\xb3\xbd",
.expected_ss =
"\x34\xc3\x35\x14\x88\x46\x26\x23\x97\xbb\xdd\x28\x5c\x94\xf6\x47"
"\xca\xb3\x19\xaf\xca\x44\x9b\xc2\x7d\x89\xfd\x96\x14\xfd\x6d\x58"
"\xd8\xc4\x6b\x61\x2a\x0d\xf2\x36\x45\xc8\xe4\xa4\xed\x81\x53\x81"
"\x66\x1e\xe0\x5a\xb1\x78\x2d\x0b\x5c\xb4\xd1\xfc\x90\xc6\x9c\xdb"
"\x5a\x30\x0b\x14\x7d\xbe\xb3\x7d\xb1\xb2\x76\x3c\x6c\xef\x74\x6b"
"\xe7\x1f\x64\x0c\xab\x65\xe1\x76\x5c\x3d\x83\xb5\x8a\xfb\xaf\x0f"
"\xf2\x06\x14\x8f\xa0\xf6\xc1\x89\x78\xf2\xba\x72\x73\x3c\xf7\x76"
"\x21\x67\xbc\x24\x31\xb8\x09\x65\x0f\x0c\x02\x32\x4a\x98\x14\xfc"
"\x72\x2c\x25\x60\x68\x5f\x2f\x30\x1e\x5b\xf0\x3b\xd1\xa2\x87\xa0"
"\x54\xdf\xdb\xc0\xee\x0a\x0f\x47\xc9\x90\x20\x2c\xf9\xe3\x52\xad"
"\x27\x65\x8d\x54\x8d\xa8\xa1\xf3\xed\x15\xd4\x94\x28\x90\x31\x93"
"\x1b\xc0\x51\xbb\x43\x5d\x76\x3b\x1d\x2a\x71\x50\xea\x5d\x48\x94"
"\x7f\x6f\xf1\x48\xdb\x30\xe5\xae\x64\x79\xd9\x7a\xdb\xc6\xff\xd8"
"\x5e\x5a\x64\xbd\xf6\x85\x04\xe8\x28\x6a\xac\xef\xce\x19\x8e\x9a"
"\xfe\x75\xc0\x27\x69\xe3\xb3\x7b\x21\xa7\xb1\x16\xa4\x85\x23\xee"
"\xb0\x1b\x04\x6e\xbd\xab\x16\xde\xfd\x86\x6b\xa9\x95\xd7\x0b\xfd",
crypto: dh - fix calculating encoded key size It was forgotten to increase DH_KPP_SECRET_MIN_SIZE to include 'q_size', causing an out-of-bounds write of 4 bytes in crypto_dh_encode_key(), and an out-of-bounds read of 4 bytes in crypto_dh_decode_key(). Fix it, and fix the lengths of the test vectors to match this. Reported-by: syzbot+6d38d558c25b53b8f4ed@syzkaller.appspotmail.com Fixes: e3fe0ae12962 ("crypto: dh - add public key verification test") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-07-27 16:36:10 -06:00
.secret_size = 533,
crypto: dh - Add DH software implementation * Implement MPI based Diffie-Hellman under kpp API * Test provided uses data generad by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:14 -06:00
.b_public_size = 256,
.expected_a_public_size = 256,
.expected_ss_size = 256,
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct kpp_testvec ecdh_tv_template[] = {
crypto: ecdh - Add ECDH software support * Implement ECDH under kpp API * Provide ECC software support for curve P-192 and P-256. * Add kpp test for ECDH with data generated by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:15 -06:00
{
#ifndef CONFIG_CRYPTO_FIPS
.secret =
#ifdef __LITTLE_ENDIAN
"\x02\x00" /* type */
"\x20\x00" /* len */
"\x01\x00" /* curve_id */
"\x18\x00" /* key_size */
#else
"\x00\x02" /* type */
"\x00\x20" /* len */
"\x00\x01" /* curve_id */
"\x00\x18" /* key_size */
#endif
"\xb5\x05\xb1\x71\x1e\xbf\x8c\xda"
"\x4e\x19\x1e\x62\x1f\x23\x23\x31"
"\x36\x1e\xd3\x84\x2f\xcc\x21\x72",
.b_public =
"\xc3\xba\x67\x4b\x71\xec\xd0\x76"
"\x7a\x99\x75\x64\x36\x13\x9a\x94"
"\x5d\x8b\xdc\x60\x90\x91\xfd\x3f"
"\xb0\x1f\x8a\x0a\x68\xc6\x88\x6e"
"\x83\x87\xdd\x67\x09\xf8\x8d\x96"
"\x07\xd6\xbd\x1c\xe6\x8d\x9d\x67",
.expected_a_public =
"\x1a\x04\xdb\xa5\xe1\xdd\x4e\x79"
"\xa3\xe6\xef\x0e\x5c\x80\x49\x85"
"\xfa\x78\xb4\xef\x49\xbd\x4c\x7c"
"\x22\x90\x21\x02\xf9\x1b\x81\x5d"
"\x0c\x8a\xa8\x98\xd6\x27\x69\x88"
"\x5e\xbc\x94\xd8\x15\x9e\x21\xce",
.expected_ss =
"\xf4\x57\xcc\x4f\x1f\x4e\x31\xcc"
"\xe3\x40\x60\xc8\x06\x93\xc6\x2e"
"\x99\x80\x81\x28\xaf\xc5\x51\x74",
.secret_size = 32,
.b_public_size = 48,
.expected_a_public_size = 48,
.expected_ss_size = 24
}, {
#endif
.secret =
#ifdef __LITTLE_ENDIAN
"\x02\x00" /* type */
"\x28\x00" /* len */
"\x02\x00" /* curve_id */
"\x20\x00" /* key_size */
#else
"\x00\x02" /* type */
"\x00\x28" /* len */
"\x00\x02" /* curve_id */
"\x00\x20" /* key_size */
#endif
"\x24\xd1\x21\xeb\xe5\xcf\x2d\x83"
"\xf6\x62\x1b\x6e\x43\x84\x3a\xa3"
"\x8b\xe0\x86\xc3\x20\x19\xda\x92"
"\x50\x53\x03\xe1\xc0\xea\xb8\x82",
.expected_a_public =
"\x1a\x7f\xeb\x52\x00\xbd\x3c\x31"
"\x7d\xb6\x70\xc1\x86\xa6\xc7\xc4"
"\x3b\xc5\x5f\x6c\x6f\x58\x3c\xf5"
"\xb6\x63\x82\x77\x33\x24\xa1\x5f"
"\x6a\xca\x43\x6f\xf7\x7e\xff\x02"
"\x37\x08\xcc\x40\x5e\x7a\xfd\x6a"
"\x6a\x02\x6e\x41\x87\x68\x38\x77"
"\xfa\xa9\x44\x43\x2d\xef\x09\xdf",
.expected_ss =
"\xea\x17\x6f\x7e\x6e\x57\x26\x38"
"\x8b\xfb\x41\xeb\xba\xc8\x6d\xa5"
"\xa8\x72\xd1\xff\xc9\x47\x3d\xaa"
"\x58\x43\x9f\x34\x0f\x8c\xf3\xc9",
.b_public =
"\xcc\xb4\xda\x74\xb1\x47\x3f\xea"
"\x6c\x70\x9e\x38\x2d\xc7\xaa\xb7"
"\x29\xb2\x47\x03\x19\xab\xdd\x34"
"\xbd\xa8\x2c\x93\xe1\xa4\x74\xd9"
"\x64\x63\xf7\x70\x20\x2f\xa4\xe6"
"\x9f\x4a\x38\xcc\xc0\x2c\x49\x2f"
"\xb1\x32\xbb\xaf\x22\x61\xda\xcb"
"\x6f\xdb\xa9\xaa\xfc\x77\x81\xf3",
.secret_size = 40,
.b_public_size = 64,
.expected_a_public_size = 64,
.expected_ss_size = 32
crypto: testmgr - add genkey kpp test The test considers a party that already has a private-public key pair and a party that provides a NULL key. The kernel will generate the private-public key pair for the latter, computes the shared secret on both ends and verifies if it's the same. The explicit private-public key pair was copied from the previous test vector. Signed-off-by: Tudor Ambarus <tudor.ambarus@microchip.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-05-30 08:52:49 -06:00
}, {
.secret =
#ifdef __LITTLE_ENDIAN
"\x02\x00" /* type */
"\x08\x00" /* len */
"\x02\x00" /* curve_id */
"\x00\x00", /* key_size */
#else
"\x00\x02" /* type */
"\x00\x08" /* len */
"\x00\x02" /* curve_id */
"\x00\x00", /* key_size */
#endif
.b_secret =
#ifdef __LITTLE_ENDIAN
"\x02\x00" /* type */
"\x28\x00" /* len */
"\x02\x00" /* curve_id */
"\x20\x00" /* key_size */
#else
"\x00\x02" /* type */
"\x00\x28" /* len */
"\x00\x02" /* curve_id */
"\x00\x20" /* key_size */
#endif
"\x24\xd1\x21\xeb\xe5\xcf\x2d\x83"
"\xf6\x62\x1b\x6e\x43\x84\x3a\xa3"
"\x8b\xe0\x86\xc3\x20\x19\xda\x92"
"\x50\x53\x03\xe1\xc0\xea\xb8\x82",
.b_public =
"\x1a\x7f\xeb\x52\x00\xbd\x3c\x31"
"\x7d\xb6\x70\xc1\x86\xa6\xc7\xc4"
"\x3b\xc5\x5f\x6c\x6f\x58\x3c\xf5"
"\xb6\x63\x82\x77\x33\x24\xa1\x5f"
"\x6a\xca\x43\x6f\xf7\x7e\xff\x02"
"\x37\x08\xcc\x40\x5e\x7a\xfd\x6a"
"\x6a\x02\x6e\x41\x87\x68\x38\x77"
"\xfa\xa9\x44\x43\x2d\xef\x09\xdf",
.secret_size = 8,
.b_secret_size = 40,
.b_public_size = 64,
.expected_a_public_size = 64,
.expected_ss_size = 32,
.genkey = true,
crypto: ecdh - Add ECDH software support * Implement ECDH under kpp API * Provide ECC software support for curve P-192 and P-256. * Add kpp test for ECDH with data generated by OpenSSL Signed-off-by: Salvatore Benedetto <salvatore.benedetto@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-22 10:49:15 -06:00
}
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* MD4 test vectors from RFC1320
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec md4_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.digest = "\x31\xd6\xcf\xe0\xd1\x6a\xe9\x31"
"\xb7\x3c\x59\xd7\xe0\xc0\x89\xc0",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\xbd\xe5\x2c\xb3\x1d\xe3\x3e\x46"
"\x24\x5e\x05\xfb\xdb\xd6\xfb\x24",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xa4\x48\x01\x7a\xaf\x21\xd8\x52"
"\x5f\xc1\x0a\xe8\x7a\xa6\x72\x9d",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\xd9\x13\x0a\x81\x64\x54\x9f\xe8"
"\x18\x87\x48\x06\xe1\xc7\x01\x4b",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xd7\x9e\x1c\x30\x8a\xa5\xbb\xcd"
"\xee\xa8\xed\x63\xdf\x41\x2d\xa9",
.np = 2,
.tap = { 13, 13 },
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\x04\x3f\x85\x82\xf2\x41\xdb\x35"
"\x1c\xe6\x27\xe1\x53\xe7\xf0\xe4",
}, {
.plaintext = "123456789012345678901234567890123456789012345678901234567890123"
"45678901234567890",
.psize = 80,
.digest = "\xe3\x3b\x4d\xdc\x9c\x38\xf2\x19"
"\x9c\x3e\x7b\x16\x4f\xcc\x05\x36",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha3_224_tv_template[] = {
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
{
.plaintext = "",
.digest = "\x6b\x4e\x03\x42\x36\x67\xdb\xb7"
"\x3b\x6e\x15\x45\x4f\x0e\xb1\xab"
"\xd4\x59\x7f\x9a\x1b\x07\x8e\x3f"
"\x5b\x5a\x6b\xc7",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x9e\x86\xff\x69\x55\x7c\xa9\x5f"
"\x40\x5f\x08\x12\x69\x68\x5b\x38"
"\xe3\xa8\x19\xb3\x09\xee\x94\x2f"
"\x48\x2b\x6a\x8b",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkl"
"jklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x8a\x24\x10\x8b\x15\x4a\xda\x21"
"\xc9\xfd\x55\x74\x49\x44\x79\xba"
"\x5c\x7e\x7a\xb7\x6e\xf2\x64\xea"
"\xd0\xfc\xce\x33",
crypto: testmgr - add new testcases for sha3 All current SHA3 test cases are smaller than the SHA3 block size, which means not all code paths are being exercised. So add a new test case to each variant, and make one of the existing test cases chunked. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-01-19 05:04:37 -07:00
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x7d\x0f\x2f\xb7\x65\x3b\xa7\x26"
"\xc3\x88\x20\x71\x15\x06\xe8\x2d"
"\xa3\x92\x44\xab\x3e\xe7\xff\x86"
"\xb6\x79\x10\x72",
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha3_256_tv_template[] = {
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
{
.plaintext = "",
.digest = "\xa7\xff\xc6\xf8\xbf\x1e\xd7\x66"
"\x51\xc1\x47\x56\xa0\x61\xd6\x62"
"\xf5\x80\xff\x4d\xe4\x3b\x49\xfa"
"\x82\xd8\x0a\x4b\x80\xf8\x43\x4a",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x80\x08\x4b\xf2\xfb\xa0\x24\x75"
"\x72\x6f\xeb\x2c\xab\x2d\x82\x15"
"\xea\xb1\x4b\xc6\xbd\xd8\xbf\xb2"
"\xc8\x15\x12\x57\x03\x2e\xcd\x8b",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkl"
"jklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x41\xc0\xdb\xa2\xa9\xd6\x24\x08"
"\x49\x10\x03\x76\xa8\x23\x5e\x2c"
"\x82\xe1\xb9\x99\x8a\x99\x9e\x21"
"\xdb\x32\xdd\x97\x49\x6d\x33\x76",
crypto: testmgr - add new testcases for sha3 All current SHA3 test cases are smaller than the SHA3 block size, which means not all code paths are being exercised. So add a new test case to each variant, and make one of the existing test cases chunked. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-01-19 05:04:37 -07:00
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\xde\x41\x04\xbd\xda\xda\xd9\x71"
"\xf7\xfa\x80\xf5\xea\x11\x03\xb1"
"\x3b\x6a\xbc\x5f\xb9\x66\x26\xf7"
"\x8a\x97\xbb\xf2\x07\x08\x38\x30",
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha3_384_tv_template[] = {
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
{
.plaintext = "",
.digest = "\x0c\x63\xa7\x5b\x84\x5e\x4f\x7d"
"\x01\x10\x7d\x85\x2e\x4c\x24\x85"
"\xc5\x1a\x50\xaa\xaa\x94\xfc\x61"
"\x99\x5e\x71\xbb\xee\x98\x3a\x2a"
"\xc3\x71\x38\x31\x26\x4a\xdb\x47"
"\xfb\x6b\xd1\xe0\x58\xd5\xf0\x04",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x18\x15\xf7\x74\xf3\x20\x49\x1b"
"\x48\x56\x9e\xfe\xc7\x94\xd2\x49"
"\xee\xb5\x9a\xae\x46\xd2\x2b\xf7"
"\x7d\xaf\xe2\x5c\x5e\xdc\x28\xd7"
"\xea\x44\xf9\x3e\xe1\x23\x4a\xa8"
"\x8f\x61\xc9\x19\x12\xa4\xcc\xd9",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkl"
"jklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x99\x1c\x66\x57\x55\xeb\x3a\x4b"
"\x6b\xbd\xfb\x75\xc7\x8a\x49\x2e"
"\x8c\x56\xa2\x2c\x5c\x4d\x7e\x42"
"\x9b\xfd\xbc\x32\xb9\xd4\xad\x5a"
"\xa0\x4a\x1f\x07\x6e\x62\xfe\xa1"
"\x9e\xef\x51\xac\xd0\x65\x7c\x22",
crypto: testmgr - add new testcases for sha3 All current SHA3 test cases are smaller than the SHA3 block size, which means not all code paths are being exercised. So add a new test case to each variant, and make one of the existing test cases chunked. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-01-19 05:04:37 -07:00
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x1b\x19\x4d\x8f\xd5\x36\x87\x71"
"\xcf\xca\x30\x85\x9b\xc1\x25\xc7"
"\x00\xcb\x73\x8a\x8e\xd4\xfe\x2b"
"\x1a\xa2\xdc\x2e\x41\xfd\x52\x51"
"\xd2\x21\xae\x2d\xc7\xae\x8c\x40"
"\xb9\xe6\x56\x48\x03\xcd\x88\x6b",
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha3_512_tv_template[] = {
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
{
.plaintext = "",
.digest = "\xa6\x9f\x73\xcc\xa2\x3a\x9a\xc5"
"\xc8\xb5\x67\xdc\x18\x5a\x75\x6e"
"\x97\xc9\x82\x16\x4f\xe2\x58\x59"
"\xe0\xd1\xdc\xc1\x47\x5c\x80\xa6"
"\x15\xb2\x12\x3a\xf1\xf5\xf9\x4c"
"\x11\xe3\xe9\x40\x2c\x3a\xc5\x58"
"\xf5\x00\x19\x9d\x95\xb6\xd3\xe3"
"\x01\x75\x85\x86\x28\x1d\xcd\x26",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x69\x7f\x2d\x85\x61\x72\xcb\x83"
"\x09\xd6\xb8\xb9\x7d\xac\x4d\xe3"
"\x44\xb5\x49\xd4\xde\xe6\x1e\xdf"
"\xb4\x96\x2d\x86\x98\xb7\xfa\x80"
"\x3f\x4f\x93\xff\x24\x39\x35\x86"
"\xe2\x8b\x5b\x95\x7a\xc3\xd1\xd3"
"\x69\x42\x0c\xe5\x33\x32\x71\x2f"
"\x99\x7b\xd3\x36\xd0\x9a\xb0\x2a",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkl"
"jklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x04\xa3\x71\xe8\x4e\xcf\xb5\xb8"
"\xb7\x7c\xb4\x86\x10\xfc\xa8\x18"
"\x2d\xd4\x57\xce\x6f\x32\x6a\x0f"
"\xd3\xd7\xec\x2f\x1e\x91\x63\x6d"
"\xee\x69\x1f\xbe\x0c\x98\x53\x02"
"\xba\x1b\x0d\x8d\xc7\x8c\x08\x63"
"\x46\xb5\x33\xb4\x9c\x03\x0d\x99"
"\xa2\x7d\xaf\x11\x39\xd6\xe7\x5e",
crypto: testmgr - add new testcases for sha3 All current SHA3 test cases are smaller than the SHA3 block size, which means not all code paths are being exercised. So add a new test case to each variant, and make one of the existing test cases chunked. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-01-19 05:04:37 -07:00
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x59\xda\x30\xe3\x90\xe4\x3d\xde"
"\xf0\xc6\x42\x17\xd7\xb2\x26\x47"
"\x90\x28\xa6\x84\xe8\x49\x7a\x86"
"\xd6\xb8\x9e\xf8\x07\x59\x21\x03"
"\xad\xd2\xed\x48\xa3\xb9\xa5\xf0"
"\xb3\xae\x02\x2b\xb8\xaf\xc3\x3b"
"\xd6\xb0\x8f\xcb\x76\x8b\xa7\x41"
"\x32\xc2\x8e\x50\x91\x86\x90\xfb",
crypto: sha3 - Add SHA-3 Test's in tcrypt Added support for SHA-3 algorithm test's in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-16 23:00:36 -06:00
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* MD5 test vectors from RFC1321
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec md5_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.digest = "\xd4\x1d\x8c\xd9\x8f\x00\xb2\x04"
"\xe9\x80\x09\x98\xec\xf8\x42\x7e",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x0c\xc1\x75\xb9\xc0\xf1\xb6\xa8"
"\x31\xc3\x99\xe2\x69\x77\x26\x61",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\x90\x01\x50\x98\x3c\xd2\x4f\xb0"
"\xd6\x96\x3f\x7d\x28\xe1\x7f\x72",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\xf9\x6b\x69\x7d\x7c\xb7\x93\x8d"
"\x52\x5a\x2f\x31\xaa\xf1\x61\xd0",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xc3\xfc\xd3\xd7\x61\x92\xe4\x00"
"\x7d\xfb\x49\x6c\xca\x67\xe1\x3b",
.np = 2,
.tap = {13, 13}
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xd1\x74\xab\x98\xd2\x77\xd9\xf5"
"\xa5\x61\x1c\x2c\x9f\x41\x9d\x9f",
}, {
.plaintext = "12345678901234567890123456789012345678901234567890123456789012"
"345678901234567890",
.psize = 80,
.digest = "\x57\xed\xf4\xa2\x2b\xe3\xc9\x55"
"\xac\x49\xda\x2e\x21\x07\xb6\x7a",
}
};
/*
* RIPEMD-128 test vectors from ISO/IEC 10118-3:2004(E)
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec rmd128_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.digest = "\xcd\xf2\x62\x13\xa1\x50\xdc\x3e"
"\xcb\x61\x0f\x18\xf6\xb3\x8b\x46",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x86\xbe\x7a\xfa\x33\x9d\x0f\xc7"
"\xcf\xc7\x85\xe7\x2f\x57\x8d\x33",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xc1\x4a\x12\x19\x9c\x66\xe4\xba"
"\x84\x63\x6b\x0f\x69\x14\x4c\x77",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x9e\x32\x7b\x3d\x6e\x52\x30\x62"
"\xaf\xc1\x13\x2d\x7d\xf9\xd1\xb8",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xfd\x2a\xa6\x07\xf7\x1d\xc8\xf5"
"\x10\x71\x49\x22\xb3\x71\x83\x4e",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcde"
"fghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xd1\xe9\x59\xeb\x17\x9c\x91\x1f"
"\xae\xa4\x62\x4c\x60\xc5\xc7\x02",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x3f\x45\xef\x19\x47\x32\xc2\xdb"
"\xb2\xc4\xa2\xc7\x69\x79\x5f\xa3",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighij"
"hijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\xa1\xaa\x06\x89\xd0\xfa\xfa\x2d"
"\xdc\x22\xe8\x8b\x49\x13\x3a\x06",
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "abcdefghbcdefghicdefghijdefghijkefghijklfghi"
"jklmghijklmnhijklmnoijklmnopjklmnopqklmnopqr"
"lmnopqrsmnopqrstnopqrstu",
.psize = 112,
.digest = "\xd4\xec\xc9\x13\xe1\xdf\x77\x6b"
"\xf4\x8d\xe9\xd5\x5b\x1f\x25\x46",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijk",
.psize = 32,
.digest = "\x13\xfc\x13\xe8\xef\xff\x34\x7d"
"\xe1\x93\xff\x46\xdb\xac\xcf\xd4",
}
};
/*
* RIPEMD-160 test vectors from ISO/IEC 10118-3:2004(E)
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec rmd160_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.digest = "\x9c\x11\x85\xa5\xc5\xe9\xfc\x54\x61\x28"
"\x08\x97\x7e\xe8\xf5\x48\xb2\x25\x8d\x31",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x0b\xdc\x9d\x2d\x25\x6b\x3e\xe9\xda\xae"
"\x34\x7b\xe6\xf4\xdc\x83\x5a\x46\x7f\xfe",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\x8e\xb2\x08\xf7\xe0\x5d\x98\x7a\x9b\x04"
"\x4a\x8e\x98\xc6\xb0\x87\xf1\x5a\x0b\xfc",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x5d\x06\x89\xef\x49\xd2\xfa\xe5\x72\xb8"
"\x81\xb1\x23\xa8\x5f\xfa\x21\x59\x5f\x36",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xf7\x1c\x27\x10\x9c\x69\x2c\x1b\x56\xbb"
"\xdc\xeb\x5b\x9d\x28\x65\xb3\x70\x8d\xbc",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcde"
"fghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xb0\xe2\x0b\x6e\x31\x16\x64\x02\x86\xed"
"\x3a\x87\xa5\x71\x30\x79\xb2\x1f\x51\x89",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x9b\x75\x2e\x45\x57\x3d\x4b\x39\xf4\xdb"
"\xd3\x32\x3c\xab\x82\xbf\x63\x32\x6b\xfb",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighij"
"hijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x12\xa0\x53\x38\x4a\x9c\x0c\x88\xe4\x05"
"\xa0\x6c\x27\xdc\xf4\x9a\xda\x62\xeb\x2b",
.np = 2,
.tap = { 28, 28 },
}, {
.plaintext = "abcdefghbcdefghicdefghijdefghijkefghijklfghi"
"jklmghijklmnhijklmnoijklmnopjklmnopqklmnopqr"
"lmnopqrsmnopqrstnopqrstu",
.psize = 112,
.digest = "\x6f\x3f\xa3\x9b\x6b\x50\x3c\x38\x4f\x91"
"\x9a\x49\xa7\xaa\x5c\x2c\x08\xbd\xfb\x45",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijk",
.psize = 32,
.digest = "\x94\xc2\x64\x11\x54\x04\xe6\x33\x79\x0d"
"\xfc\xc8\x7b\x58\x7d\x36\x77\x06\x7d\x9f",
}
};
/*
* RIPEMD-256 test vectors
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec rmd256_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.digest = "\x02\xba\x4c\x4e\x5f\x8e\xcd\x18"
"\x77\xfc\x52\xd6\x4d\x30\xe3\x7a"
"\x2d\x97\x74\xfb\x1e\x5d\x02\x63"
"\x80\xae\x01\x68\xe3\xc5\x52\x2d",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\xf9\x33\x3e\x45\xd8\x57\xf5\xd9"
"\x0a\x91\xba\xb7\x0a\x1e\xba\x0c"
"\xfb\x1b\xe4\xb0\x78\x3c\x9a\xcf"
"\xcd\x88\x3a\x91\x34\x69\x29\x25",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xaf\xbd\x6e\x22\x8b\x9d\x8c\xbb"
"\xce\xf5\xca\x2d\x03\xe6\xdb\xa1"
"\x0a\xc0\xbc\x7d\xcb\xe4\x68\x0e"
"\x1e\x42\xd2\xe9\x75\x45\x9b\x65",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x87\xe9\x71\x75\x9a\x1c\xe4\x7a"
"\x51\x4d\x5c\x91\x4c\x39\x2c\x90"
"\x18\xc7\xc4\x6b\xc1\x44\x65\x55"
"\x4a\xfc\xdf\x54\xa5\x07\x0c\x0e",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\x64\x9d\x30\x34\x75\x1e\xa2\x16"
"\x77\x6b\xf9\xa1\x8a\xcc\x81\xbc"
"\x78\x96\x11\x8a\x51\x97\x96\x87"
"\x82\xdd\x1f\xd9\x7d\x8d\x51\x33",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcde"
"fghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\x57\x40\xa4\x08\xac\x16\xb7\x20"
"\xb8\x44\x24\xae\x93\x1c\xbb\x1f"
"\xe3\x63\xd1\xd0\xbf\x40\x17\xf1"
"\xa8\x9f\x7e\xa6\xde\x77\xa0\xb8",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x06\xfd\xcc\x7a\x40\x95\x48\xaa"
"\xf9\x13\x68\xc0\x6a\x62\x75\xb5"
"\x53\xe3\xf0\x99\xbf\x0e\xa4\xed"
"\xfd\x67\x78\xdf\x89\xa8\x90\xdd",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighij"
"hijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x38\x43\x04\x55\x83\xaa\xc6\xc8"
"\xc8\xd9\x12\x85\x73\xe7\xa9\x80"
"\x9a\xfb\x2a\x0f\x34\xcc\xc3\x6e"
"\xa9\xe7\x2f\x16\xf6\x36\x8e\x3f",
.np = 2,
.tap = { 28, 28 },
}
};
/*
* RIPEMD-320 test vectors
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec rmd320_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.digest = "\x22\xd6\x5d\x56\x61\x53\x6c\xdc\x75\xc1"
"\xfd\xf5\xc6\xde\x7b\x41\xb9\xf2\x73\x25"
"\xeb\xc6\x1e\x85\x57\x17\x7d\x70\x5a\x0e"
"\xc8\x80\x15\x1c\x3a\x32\xa0\x08\x99\xb8",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\xce\x78\x85\x06\x38\xf9\x26\x58\xa5\xa5"
"\x85\x09\x75\x79\x92\x6d\xda\x66\x7a\x57"
"\x16\x56\x2c\xfc\xf6\xfb\xe7\x7f\x63\x54"
"\x2f\x99\xb0\x47\x05\xd6\x97\x0d\xff\x5d",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xde\x4c\x01\xb3\x05\x4f\x89\x30\xa7\x9d"
"\x09\xae\x73\x8e\x92\x30\x1e\x5a\x17\x08"
"\x5b\xef\xfd\xc1\xb8\xd1\x16\x71\x3e\x74"
"\xf8\x2f\xa9\x42\xd6\x4c\xdb\xc4\x68\x2d",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x3a\x8e\x28\x50\x2e\xd4\x5d\x42\x2f\x68"
"\x84\x4f\x9d\xd3\x16\xe7\xb9\x85\x33\xfa"
"\x3f\x2a\x91\xd2\x9f\x84\xd4\x25\xc8\x8d"
"\x6b\x4e\xff\x72\x7d\xf6\x6a\x7c\x01\x97",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xca\xbd\xb1\x81\x0b\x92\x47\x0a\x20\x93"
"\xaa\x6b\xce\x05\x95\x2c\x28\x34\x8c\xf4"
"\x3f\xf6\x08\x41\x97\x51\x66\xbb\x40\xed"
"\x23\x40\x04\xb8\x82\x44\x63\xe6\xb0\x09",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcde"
"fghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xed\x54\x49\x40\xc8\x6d\x67\xf2\x50\xd2"
"\x32\xc3\x0b\x7b\x3e\x57\x70\xe0\xc6\x0c"
"\x8c\xb9\xa4\xca\xfe\x3b\x11\x38\x8a\xf9"
"\x92\x0e\x1b\x99\x23\x0b\x84\x3c\x86\xa4",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x55\x78\x88\xaf\x5f\x6d\x8e\xd6\x2a\xb6"
"\x69\x45\xc6\xd2\xa0\xa4\x7e\xcd\x53\x41"
"\xe9\x15\xeb\x8f\xea\x1d\x05\x24\x95\x5f"
"\x82\x5d\xc7\x17\xe4\xa0\x08\xab\x2d\x42",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighij"
"hijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\xd0\x34\xa7\x95\x0c\xf7\x22\x02\x1b\xa4"
"\xb8\x4d\xf7\x69\xa5\xde\x20\x60\xe2\x59"
"\xdf\x4c\x9b\xb4\xa4\x26\x8c\x0e\x93\x5b"
"\xbc\x74\x70\xa9\x69\xc9\xd0\x72\xa1\xac",
.np = 2,
.tap = { 28, 28 },
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec crct10dif_tv_template[] = {
Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework" This patch reinstates commits 67822649d7305caf3dd50ed46c27b99c94eff996 39761214eefc6b070f29402aa1165f24d789b3f7 0b95a7f85718adcbba36407ef88bba0a7379ed03 31d939625a9a20b1badd2d4e6bf6fd39fa523405 2d31e518a42828df7877bca23a958627d60408bc Now that module softdeps are in the kernel we can use that to resolve the boot issue which cause the revert. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-06 20:56:26 -06:00
{
crypto: testmgr - add/enhance test cases for CRC-T10DIF The existing test cases only exercise a small slice of the various possible code paths through the x86 SSE/PCLMULQDQ implementation, and the upcoming ports of it for arm64. So add one that exceeds 256 bytes in size, and convert another to a chunked test. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-12-05 11:42:24 -07:00
.plaintext = "abc",
.psize = 3,
.digest = (u8 *)(u16 []){ 0x443b },
Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework" This patch reinstates commits 67822649d7305caf3dd50ed46c27b99c94eff996 39761214eefc6b070f29402aa1165f24d789b3f7 0b95a7f85718adcbba36407ef88bba0a7379ed03 31d939625a9a20b1badd2d4e6bf6fd39fa523405 2d31e518a42828df7877bca23a958627d60408bc Now that module softdeps are in the kernel we can use that to resolve the boot issue which cause the revert. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-06 20:56:26 -06:00
}, {
crypto: testmgr - add/enhance test cases for CRC-T10DIF The existing test cases only exercise a small slice of the various possible code paths through the x86 SSE/PCLMULQDQ implementation, and the upcoming ports of it for arm64. So add one that exceeds 256 bytes in size, and convert another to a chunked test. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-12-05 11:42:24 -07:00
.plaintext = "1234567890123456789012345678901234567890"
"123456789012345678901234567890123456789",
.psize = 79,
.digest = (u8 *)(u16 []){ 0x4b70 },
.np = 2,
.tap = { 63, 16 },
Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework" This patch reinstates commits 67822649d7305caf3dd50ed46c27b99c94eff996 39761214eefc6b070f29402aa1165f24d789b3f7 0b95a7f85718adcbba36407ef88bba0a7379ed03 31d939625a9a20b1badd2d4e6bf6fd39fa523405 2d31e518a42828df7877bca23a958627d60408bc Now that module softdeps are in the kernel we can use that to resolve the boot issue which cause the revert. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-06 20:56:26 -06:00
}, {
crypto: testmgr - add/enhance test cases for CRC-T10DIF The existing test cases only exercise a small slice of the various possible code paths through the x86 SSE/PCLMULQDQ implementation, and the upcoming ports of it for arm64. So add one that exceeds 256 bytes in size, and convert another to a chunked test. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-12-05 11:42:24 -07:00
.plaintext = "abcdddddddddddddddddddddddddddddddddddddddd"
"ddddddddddddd",
.psize = 56,
.digest = (u8 *)(u16 []){ 0x9ce3 },
.np = 8,
.tap = { 1, 2, 28, 7, 6, 5, 4, 3 },
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"123456789012345678901234567890123456789",
.psize = 319,
.digest = (u8 *)(u16 []){ 0x44c6 },
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890"
"123456789012345678901234567890123456789",
.psize = 319,
.digest = (u8 *)(u16 []){ 0x44c6 },
.np = 4,
.tap = { 1, 255, 57, 6 },
crypto: testmgr - add a new test case for CRC-T10DIF In order to be able to test yield support under preempt, add a test vector for CRC-T10DIF that is long enough to take multiple iterations (and thus possible preemption between them) of the primary loop of the accelerated x86 and arm64 implementations. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-03-10 08:21:46 -07:00
}, {
.plaintext = "\x6e\x05\x79\x10\xa7\x1b\xb2\x49"
"\xe0\x54\xeb\x82\x19\x8d\x24\xbb"
"\x2f\xc6\x5d\xf4\x68\xff\x96\x0a"
"\xa1\x38\xcf\x43\xda\x71\x08\x7c"
"\x13\xaa\x1e\xb5\x4c\xe3\x57\xee"
"\x85\x1c\x90\x27\xbe\x32\xc9\x60"
"\xf7\x6b\x02\x99\x0d\xa4\x3b\xd2"
"\x46\xdd\x74\x0b\x7f\x16\xad\x21"
"\xb8\x4f\xe6\x5a\xf1\x88\x1f\x93"
"\x2a\xc1\x35\xcc\x63\xfa\x6e\x05"
"\x9c\x10\xa7\x3e\xd5\x49\xe0\x77"
"\x0e\x82\x19\xb0\x24\xbb\x52\xe9"
"\x5d\xf4\x8b\x22\x96\x2d\xc4\x38"
"\xcf\x66\xfd\x71\x08\x9f\x13\xaa"
"\x41\xd8\x4c\xe3\x7a\x11\x85\x1c"
"\xb3\x27\xbe\x55\xec\x60\xf7\x8e"
"\x02\x99\x30\xc7\x3b\xd2\x69\x00"
"\x74\x0b\xa2\x16\xad\x44\xdb\x4f"
"\xe6\x7d\x14\x88\x1f\xb6\x2a\xc1"
"\x58\xef\x63\xfa\x91\x05\x9c\x33"
"\xca\x3e\xd5\x6c\x03\x77\x0e\xa5"
"\x19\xb0\x47\xde\x52\xe9\x80\x17"
"\x8b\x22\xb9\x2d\xc4\x5b\xf2\x66"
"\xfd\x94\x08\x9f\x36\xcd\x41\xd8"
"\x6f\x06\x7a\x11\xa8\x1c\xb3\x4a"
"\xe1\x55\xec\x83\x1a\x8e\x25\xbc"
"\x30\xc7\x5e\xf5\x69\x00\x97\x0b"
"\xa2\x39\xd0\x44\xdb\x72\x09\x7d"
"\x14\xab\x1f\xb6\x4d\xe4\x58\xef"
"\x86\x1d\x91\x28\xbf\x33\xca\x61"
"\xf8\x6c\x03\x9a\x0e\xa5\x3c\xd3"
"\x47\xde\x75\x0c\x80\x17\xae\x22"
"\xb9\x50\xe7\x5b\xf2\x89\x20\x94"
"\x2b\xc2\x36\xcd\x64\xfb\x6f\x06"
"\x9d\x11\xa8\x3f\xd6\x4a\xe1\x78"
"\x0f\x83\x1a\xb1\x25\xbc\x53\xea"
"\x5e\xf5\x8c\x00\x97\x2e\xc5\x39"
"\xd0\x67\xfe\x72\x09\xa0\x14\xab"
"\x42\xd9\x4d\xe4\x7b\x12\x86\x1d"
"\xb4\x28\xbf\x56\xed\x61\xf8\x8f"
"\x03\x9a\x31\xc8\x3c\xd3\x6a\x01"
"\x75\x0c\xa3\x17\xae\x45\xdc\x50"
"\xe7\x7e\x15\x89\x20\xb7\x2b\xc2"
"\x59\xf0\x64\xfb\x92\x06\x9d\x34"
"\xcb\x3f\xd6\x6d\x04\x78\x0f\xa6"
"\x1a\xb1\x48\xdf\x53\xea\x81\x18"
"\x8c\x23\xba\x2e\xc5\x5c\xf3\x67"
"\xfe\x95\x09\xa0\x37\xce\x42\xd9"
"\x70\x07\x7b\x12\xa9\x1d\xb4\x4b"
"\xe2\x56\xed\x84\x1b\x8f\x26\xbd"
"\x31\xc8\x5f\xf6\x6a\x01\x98\x0c"
"\xa3\x3a\xd1\x45\xdc\x73\x0a\x7e"
"\x15\xac\x20\xb7\x4e\xe5\x59\xf0"
"\x87\x1e\x92\x29\xc0\x34\xcb\x62"
"\xf9\x6d\x04\x9b\x0f\xa6\x3d\xd4"
"\x48\xdf\x76\x0d\x81\x18\xaf\x23"
"\xba\x51\xe8\x5c\xf3\x8a\x21\x95"
"\x2c\xc3\x37\xce\x65\xfc\x70\x07"
"\x9e\x12\xa9\x40\xd7\x4b\xe2\x79"
"\x10\x84\x1b\xb2\x26\xbd\x54\xeb"
"\x5f\xf6\x8d\x01\x98\x2f\xc6\x3a"
"\xd1\x68\xff\x73\x0a\xa1\x15\xac"
"\x43\xda\x4e\xe5\x7c\x13\x87\x1e"
"\xb5\x29\xc0\x57\xee\x62\xf9\x90"
"\x04\x9b\x32\xc9\x3d\xd4\x6b\x02"
"\x76\x0d\xa4\x18\xaf\x46\xdd\x51"
"\xe8\x7f\x16\x8a\x21\xb8\x2c\xc3"
"\x5a\xf1\x65\xfc\x93\x07\x9e\x35"
"\xcc\x40\xd7\x6e\x05\x79\x10\xa7"
"\x1b\xb2\x49\xe0\x54\xeb\x82\x19"
"\x8d\x24\xbb\x2f\xc6\x5d\xf4\x68"
"\xff\x96\x0a\xa1\x38\xcf\x43\xda"
"\x71\x08\x7c\x13\xaa\x1e\xb5\x4c"
"\xe3\x57\xee\x85\x1c\x90\x27\xbe"
"\x32\xc9\x60\xf7\x6b\x02\x99\x0d"
"\xa4\x3b\xd2\x46\xdd\x74\x0b\x7f"
"\x16\xad\x21\xb8\x4f\xe6\x5a\xf1"
"\x88\x1f\x93\x2a\xc1\x35\xcc\x63"
"\xfa\x6e\x05\x9c\x10\xa7\x3e\xd5"
"\x49\xe0\x77\x0e\x82\x19\xb0\x24"
"\xbb\x52\xe9\x5d\xf4\x8b\x22\x96"
"\x2d\xc4\x38\xcf\x66\xfd\x71\x08"
"\x9f\x13\xaa\x41\xd8\x4c\xe3\x7a"
"\x11\x85\x1c\xb3\x27\xbe\x55\xec"
"\x60\xf7\x8e\x02\x99\x30\xc7\x3b"
"\xd2\x69\x00\x74\x0b\xa2\x16\xad"
"\x44\xdb\x4f\xe6\x7d\x14\x88\x1f"
"\xb6\x2a\xc1\x58\xef\x63\xfa\x91"
"\x05\x9c\x33\xca\x3e\xd5\x6c\x03"
"\x77\x0e\xa5\x19\xb0\x47\xde\x52"
"\xe9\x80\x17\x8b\x22\xb9\x2d\xc4"
"\x5b\xf2\x66\xfd\x94\x08\x9f\x36"
"\xcd\x41\xd8\x6f\x06\x7a\x11\xa8"
"\x1c\xb3\x4a\xe1\x55\xec\x83\x1a"
"\x8e\x25\xbc\x30\xc7\x5e\xf5\x69"
"\x00\x97\x0b\xa2\x39\xd0\x44\xdb"
"\x72\x09\x7d\x14\xab\x1f\xb6\x4d"
"\xe4\x58\xef\x86\x1d\x91\x28\xbf"
"\x33\xca\x61\xf8\x6c\x03\x9a\x0e"
"\xa5\x3c\xd3\x47\xde\x75\x0c\x80"
"\x17\xae\x22\xb9\x50\xe7\x5b\xf2"
"\x89\x20\x94\x2b\xc2\x36\xcd\x64"
"\xfb\x6f\x06\x9d\x11\xa8\x3f\xd6"
"\x4a\xe1\x78\x0f\x83\x1a\xb1\x25"
"\xbc\x53\xea\x5e\xf5\x8c\x00\x97"
"\x2e\xc5\x39\xd0\x67\xfe\x72\x09"
"\xa0\x14\xab\x42\xd9\x4d\xe4\x7b"
"\x12\x86\x1d\xb4\x28\xbf\x56\xed"
"\x61\xf8\x8f\x03\x9a\x31\xc8\x3c"
"\xd3\x6a\x01\x75\x0c\xa3\x17\xae"
"\x45\xdc\x50\xe7\x7e\x15\x89\x20"
"\xb7\x2b\xc2\x59\xf0\x64\xfb\x92"
"\x06\x9d\x34\xcb\x3f\xd6\x6d\x04"
"\x78\x0f\xa6\x1a\xb1\x48\xdf\x53"
"\xea\x81\x18\x8c\x23\xba\x2e\xc5"
"\x5c\xf3\x67\xfe\x95\x09\xa0\x37"
"\xce\x42\xd9\x70\x07\x7b\x12\xa9"
"\x1d\xb4\x4b\xe2\x56\xed\x84\x1b"
"\x8f\x26\xbd\x31\xc8\x5f\xf6\x6a"
"\x01\x98\x0c\xa3\x3a\xd1\x45\xdc"
"\x73\x0a\x7e\x15\xac\x20\xb7\x4e"
"\xe5\x59\xf0\x87\x1e\x92\x29\xc0"
"\x34\xcb\x62\xf9\x6d\x04\x9b\x0f"
"\xa6\x3d\xd4\x48\xdf\x76\x0d\x81"
"\x18\xaf\x23\xba\x51\xe8\x5c\xf3"
"\x8a\x21\x95\x2c\xc3\x37\xce\x65"
"\xfc\x70\x07\x9e\x12\xa9\x40\xd7"
"\x4b\xe2\x79\x10\x84\x1b\xb2\x26"
"\xbd\x54\xeb\x5f\xf6\x8d\x01\x98"
"\x2f\xc6\x3a\xd1\x68\xff\x73\x0a"
"\xa1\x15\xac\x43\xda\x4e\xe5\x7c"
"\x13\x87\x1e\xb5\x29\xc0\x57\xee"
"\x62\xf9\x90\x04\x9b\x32\xc9\x3d"
"\xd4\x6b\x02\x76\x0d\xa4\x18\xaf"
"\x46\xdd\x51\xe8\x7f\x16\x8a\x21"
"\xb8\x2c\xc3\x5a\xf1\x65\xfc\x93"
"\x07\x9e\x35\xcc\x40\xd7\x6e\x05"
"\x79\x10\xa7\x1b\xb2\x49\xe0\x54"
"\xeb\x82\x19\x8d\x24\xbb\x2f\xc6"
"\x5d\xf4\x68\xff\x96\x0a\xa1\x38"
"\xcf\x43\xda\x71\x08\x7c\x13\xaa"
"\x1e\xb5\x4c\xe3\x57\xee\x85\x1c"
"\x90\x27\xbe\x32\xc9\x60\xf7\x6b"
"\x02\x99\x0d\xa4\x3b\xd2\x46\xdd"
"\x74\x0b\x7f\x16\xad\x21\xb8\x4f"
"\xe6\x5a\xf1\x88\x1f\x93\x2a\xc1"
"\x35\xcc\x63\xfa\x6e\x05\x9c\x10"
"\xa7\x3e\xd5\x49\xe0\x77\x0e\x82"
"\x19\xb0\x24\xbb\x52\xe9\x5d\xf4"
"\x8b\x22\x96\x2d\xc4\x38\xcf\x66"
"\xfd\x71\x08\x9f\x13\xaa\x41\xd8"
"\x4c\xe3\x7a\x11\x85\x1c\xb3\x27"
"\xbe\x55\xec\x60\xf7\x8e\x02\x99"
"\x30\xc7\x3b\xd2\x69\x00\x74\x0b"
"\xa2\x16\xad\x44\xdb\x4f\xe6\x7d"
"\x14\x88\x1f\xb6\x2a\xc1\x58\xef"
"\x63\xfa\x91\x05\x9c\x33\xca\x3e"
"\xd5\x6c\x03\x77\x0e\xa5\x19\xb0"
"\x47\xde\x52\xe9\x80\x17\x8b\x22"
"\xb9\x2d\xc4\x5b\xf2\x66\xfd\x94"
"\x08\x9f\x36\xcd\x41\xd8\x6f\x06"
"\x7a\x11\xa8\x1c\xb3\x4a\xe1\x55"
"\xec\x83\x1a\x8e\x25\xbc\x30\xc7"
"\x5e\xf5\x69\x00\x97\x0b\xa2\x39"
"\xd0\x44\xdb\x72\x09\x7d\x14\xab"
"\x1f\xb6\x4d\xe4\x58\xef\x86\x1d"
"\x91\x28\xbf\x33\xca\x61\xf8\x6c"
"\x03\x9a\x0e\xa5\x3c\xd3\x47\xde"
"\x75\x0c\x80\x17\xae\x22\xb9\x50"
"\xe7\x5b\xf2\x89\x20\x94\x2b\xc2"
"\x36\xcd\x64\xfb\x6f\x06\x9d\x11"
"\xa8\x3f\xd6\x4a\xe1\x78\x0f\x83"
"\x1a\xb1\x25\xbc\x53\xea\x5e\xf5"
"\x8c\x00\x97\x2e\xc5\x39\xd0\x67"
"\xfe\x72\x09\xa0\x14\xab\x42\xd9"
"\x4d\xe4\x7b\x12\x86\x1d\xb4\x28"
"\xbf\x56\xed\x61\xf8\x8f\x03\x9a"
"\x31\xc8\x3c\xd3\x6a\x01\x75\x0c"
"\xa3\x17\xae\x45\xdc\x50\xe7\x7e"
"\x15\x89\x20\xb7\x2b\xc2\x59\xf0"
"\x64\xfb\x92\x06\x9d\x34\xcb\x3f"
"\xd6\x6d\x04\x78\x0f\xa6\x1a\xb1"
"\x48\xdf\x53\xea\x81\x18\x8c\x23"
"\xba\x2e\xc5\x5c\xf3\x67\xfe\x95"
"\x09\xa0\x37\xce\x42\xd9\x70\x07"
"\x7b\x12\xa9\x1d\xb4\x4b\xe2\x56"
"\xed\x84\x1b\x8f\x26\xbd\x31\xc8"
"\x5f\xf6\x6a\x01\x98\x0c\xa3\x3a"
"\xd1\x45\xdc\x73\x0a\x7e\x15\xac"
"\x20\xb7\x4e\xe5\x59\xf0\x87\x1e"
"\x92\x29\xc0\x34\xcb\x62\xf9\x6d"
"\x04\x9b\x0f\xa6\x3d\xd4\x48\xdf"
"\x76\x0d\x81\x18\xaf\x23\xba\x51"
"\xe8\x5c\xf3\x8a\x21\x95\x2c\xc3"
"\x37\xce\x65\xfc\x70\x07\x9e\x12"
"\xa9\x40\xd7\x4b\xe2\x79\x10\x84"
"\x1b\xb2\x26\xbd\x54\xeb\x5f\xf6"
"\x8d\x01\x98\x2f\xc6\x3a\xd1\x68"
"\xff\x73\x0a\xa1\x15\xac\x43\xda"
"\x4e\xe5\x7c\x13\x87\x1e\xb5\x29"
"\xc0\x57\xee\x62\xf9\x90\x04\x9b"
"\x32\xc9\x3d\xd4\x6b\x02\x76\x0d"
"\xa4\x18\xaf\x46\xdd\x51\xe8\x7f"
"\x16\x8a\x21\xb8\x2c\xc3\x5a\xf1"
"\x65\xfc\x93\x07\x9e\x35\xcc\x40"
"\xd7\x6e\x05\x79\x10\xa7\x1b\xb2"
"\x49\xe0\x54\xeb\x82\x19\x8d\x24"
"\xbb\x2f\xc6\x5d\xf4\x68\xff\x96"
"\x0a\xa1\x38\xcf\x43\xda\x71\x08"
"\x7c\x13\xaa\x1e\xb5\x4c\xe3\x57"
"\xee\x85\x1c\x90\x27\xbe\x32\xc9"
"\x60\xf7\x6b\x02\x99\x0d\xa4\x3b"
"\xd2\x46\xdd\x74\x0b\x7f\x16\xad"
"\x21\xb8\x4f\xe6\x5a\xf1\x88\x1f"
"\x93\x2a\xc1\x35\xcc\x63\xfa\x6e"
"\x05\x9c\x10\xa7\x3e\xd5\x49\xe0"
"\x77\x0e\x82\x19\xb0\x24\xbb\x52"
"\xe9\x5d\xf4\x8b\x22\x96\x2d\xc4"
"\x38\xcf\x66\xfd\x71\x08\x9f\x13"
"\xaa\x41\xd8\x4c\xe3\x7a\x11\x85"
"\x1c\xb3\x27\xbe\x55\xec\x60\xf7"
"\x8e\x02\x99\x30\xc7\x3b\xd2\x69"
"\x00\x74\x0b\xa2\x16\xad\x44\xdb"
"\x4f\xe6\x7d\x14\x88\x1f\xb6\x2a"
"\xc1\x58\xef\x63\xfa\x91\x05\x9c"
"\x33\xca\x3e\xd5\x6c\x03\x77\x0e"
"\xa5\x19\xb0\x47\xde\x52\xe9\x80"
"\x17\x8b\x22\xb9\x2d\xc4\x5b\xf2"
"\x66\xfd\x94\x08\x9f\x36\xcd\x41"
"\xd8\x6f\x06\x7a\x11\xa8\x1c\xb3"
"\x4a\xe1\x55\xec\x83\x1a\x8e\x25"
"\xbc\x30\xc7\x5e\xf5\x69\x00\x97"
"\x0b\xa2\x39\xd0\x44\xdb\x72\x09"
"\x7d\x14\xab\x1f\xb6\x4d\xe4\x58"
"\xef\x86\x1d\x91\x28\xbf\x33\xca"
"\x61\xf8\x6c\x03\x9a\x0e\xa5\x3c"
"\xd3\x47\xde\x75\x0c\x80\x17\xae"
"\x22\xb9\x50\xe7\x5b\xf2\x89\x20"
"\x94\x2b\xc2\x36\xcd\x64\xfb\x6f"
"\x06\x9d\x11\xa8\x3f\xd6\x4a\xe1"
"\x78\x0f\x83\x1a\xb1\x25\xbc\x53"
"\xea\x5e\xf5\x8c\x00\x97\x2e\xc5"
"\x39\xd0\x67\xfe\x72\x09\xa0\x14"
"\xab\x42\xd9\x4d\xe4\x7b\x12\x86"
"\x1d\xb4\x28\xbf\x56\xed\x61\xf8"
"\x8f\x03\x9a\x31\xc8\x3c\xd3\x6a"
"\x01\x75\x0c\xa3\x17\xae\x45\xdc"
"\x50\xe7\x7e\x15\x89\x20\xb7\x2b"
"\xc2\x59\xf0\x64\xfb\x92\x06\x9d"
"\x34\xcb\x3f\xd6\x6d\x04\x78\x0f"
"\xa6\x1a\xb1\x48\xdf\x53\xea\x81"
"\x18\x8c\x23\xba\x2e\xc5\x5c\xf3"
"\x67\xfe\x95\x09\xa0\x37\xce\x42"
"\xd9\x70\x07\x7b\x12\xa9\x1d\xb4"
"\x4b\xe2\x56\xed\x84\x1b\x8f\x26"
"\xbd\x31\xc8\x5f\xf6\x6a\x01\x98",
.psize = 2048,
.digest = (u8 *)(u16 []){ 0x23ca },
Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework" This patch reinstates commits 67822649d7305caf3dd50ed46c27b99c94eff996 39761214eefc6b070f29402aa1165f24d789b3f7 0b95a7f85718adcbba36407ef88bba0a7379ed03 31d939625a9a20b1badd2d4e6bf6fd39fa523405 2d31e518a42828df7877bca23a958627d60408bc Now that module softdeps are in the kernel we can use that to resolve the boot issue which cause the revert. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-06 20:56:26 -06:00
}
crypto: sm3 - add SM3 test vectors Add testmgr and tcrypt tests and vectors for SM3 secure hash. Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-08-21 04:51:29 -06:00
};
/* Example vectors below taken from
* http://www.oscca.gov.cn/UpFile/20101222141857786.pdf
*
* The rest taken from
* https://github.com/adamws/oscca-sm3
*/
static const struct hash_testvec sm3_tv_template[] = {
{
.plaintext = "",
.psize = 0,
.digest = (u8 *)(u8 []) {
0x1A, 0xB2, 0x1D, 0x83, 0x55, 0xCF, 0xA1, 0x7F,
0x8e, 0x61, 0x19, 0x48, 0x31, 0xE8, 0x1A, 0x8F,
0x22, 0xBE, 0xC8, 0xC7, 0x28, 0xFE, 0xFB, 0x74,
0x7E, 0xD0, 0x35, 0xEB, 0x50, 0x82, 0xAA, 0x2B }
}, {
.plaintext = "a",
.psize = 1,
.digest = (u8 *)(u8 []) {
0x62, 0x34, 0x76, 0xAC, 0x18, 0xF6, 0x5A, 0x29,
0x09, 0xE4, 0x3C, 0x7F, 0xEC, 0x61, 0xB4, 0x9C,
0x7E, 0x76, 0x4A, 0x91, 0xA1, 0x8C, 0xCB, 0x82,
0xF1, 0x91, 0x7A, 0x29, 0xC8, 0x6C, 0x5E, 0x88 }
}, {
/* A.1. Example 1 */
.plaintext = "abc",
.psize = 3,
.digest = (u8 *)(u8 []) {
0x66, 0xC7, 0xF0, 0xF4, 0x62, 0xEE, 0xED, 0xD9,
0xD1, 0xF2, 0xD4, 0x6B, 0xDC, 0x10, 0xE4, 0xE2,
0x41, 0x67, 0xC4, 0x87, 0x5C, 0xF2, 0xF7, 0xA2,
0x29, 0x7D, 0xA0, 0x2B, 0x8F, 0x4B, 0xA8, 0xE0 }
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = (u8 *)(u8 []) {
0xB8, 0x0F, 0xE9, 0x7A, 0x4D, 0xA2, 0x4A, 0xFC,
0x27, 0x75, 0x64, 0xF6, 0x6A, 0x35, 0x9E, 0xF4,
0x40, 0x46, 0x2A, 0xD2, 0x8D, 0xCC, 0x6D, 0x63,
0xAD, 0xB2, 0x4D, 0x5C, 0x20, 0xA6, 0x15, 0x95 }
}, {
/* A.1. Example 2 */
.plaintext = "abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdab"
"cdabcdabcdabcdabcd",
.psize = 64,
.digest = (u8 *)(u8 []) {
0xDE, 0xBE, 0x9F, 0xF9, 0x22, 0x75, 0xB8, 0xA1,
0x38, 0x60, 0x48, 0x89, 0xC1, 0x8E, 0x5A, 0x4D,
0x6F, 0xDB, 0x70, 0xE5, 0x38, 0x7E, 0x57, 0x65,
0x29, 0x3D, 0xCB, 0xA3, 0x9C, 0x0C, 0x57, 0x32 }
}, {
.plaintext = "abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcdabcd"
"abcdabcdabcdabcdabcdabcdabcdabcd",
.psize = 256,
.digest = (u8 *)(u8 []) {
0xB9, 0x65, 0x76, 0x4C, 0x8B, 0xEB, 0xB0, 0x91,
0xC7, 0x60, 0x2B, 0x74, 0xAF, 0xD3, 0x4E, 0xEF,
0xB5, 0x31, 0xDC, 0xCB, 0x4E, 0x00, 0x76, 0xD9,
0xB7, 0xCD, 0x81, 0x31, 0x99, 0xB4, 0x59, 0x71 }
}
Reinstate "crypto: crct10dif - Wrap crc_t10dif function all to use crypto transform framework" This patch reinstates commits 67822649d7305caf3dd50ed46c27b99c94eff996 39761214eefc6b070f29402aa1165f24d789b3f7 0b95a7f85718adcbba36407ef88bba0a7379ed03 31d939625a9a20b1badd2d4e6bf6fd39fa523405 2d31e518a42828df7877bca23a958627d60408bc Now that module softdeps are in the kernel we can use that to resolve the boot issue which cause the revert. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-09-06 20:56:26 -06:00
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* SHA1 test vectors from from FIPS PUB 180-1
crypto: sha1 - Add test vector to test partial block processing In light of the recent discovery of the bug with partial block processing on s390, we need best test coverage for that. This patch adds a test vector for SHA1 that should catch such problems. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-02-16 20:24:45 -07:00
* Long vector from CAVS 5.0
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha1_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
.plaintext = "",
.psize = 0,
.digest = "\xda\x39\xa3\xee\x5e\x6b\x4b\x0d\x32\x55"
"\xbf\xef\x95\x60\x18\x90\xaf\xd8\x07\x09",
}, {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.plaintext = "abc",
.psize = 3,
.digest = "\xa9\x99\x3e\x36\x47\x06\x81\x6a\xba\x3e"
"\x25\x71\x78\x50\xc2\x6c\x9c\xd0\xd8\x9d",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x84\x98\x3e\x44\x1c\x3b\xd2\x6e\xba\xae"
"\x4a\xa1\xf9\x51\x29\xe5\xe5\x46\x70\xf1",
.np = 2,
.tap = { 28, 28 }
crypto: sha1 - Add test vector to test partial block processing In light of the recent discovery of the bug with partial block processing on s390, we need best test coverage for that. This patch adds a test vector for SHA1 that should catch such problems. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-02-16 20:24:45 -07:00
}, {
.plaintext = "\xec\x29\x56\x12\x44\xed\xe7\x06"
"\xb6\xeb\x30\xa1\xc3\x71\xd7\x44"
"\x50\xa1\x05\xc3\xf9\x73\x5f\x7f"
"\xa9\xfe\x38\xcf\x67\xf3\x04\xa5"
"\x73\x6a\x10\x6e\x92\xe1\x71\x39"
"\xa6\x81\x3b\x1c\x81\xa4\xf3\xd3"
"\xfb\x95\x46\xab\x42\x96\xfa\x9f"
"\x72\x28\x26\xc0\x66\x86\x9e\xda"
"\xcd\x73\xb2\x54\x80\x35\x18\x58"
"\x13\xe2\x26\x34\xa9\xda\x44\x00"
"\x0d\x95\xa2\x81\xff\x9f\x26\x4e"
"\xcc\xe0\xa9\x31\x22\x21\x62\xd0"
"\x21\xcc\xa2\x8d\xb5\xf3\xc2\xaa"
"\x24\x94\x5a\xb1\xe3\x1c\xb4\x13"
"\xae\x29\x81\x0f\xd7\x94\xca\xd5"
"\xdf\xaf\x29\xec\x43\xcb\x38\xd1"
"\x98\xfe\x4a\xe1\xda\x23\x59\x78"
"\x02\x21\x40\x5b\xd6\x71\x2a\x53"
"\x05\xda\x4b\x1b\x73\x7f\xce\x7c"
"\xd2\x1c\x0e\xb7\x72\x8d\x08\x23"
"\x5a\x90\x11",
.psize = 163,
.digest = "\x97\x01\x11\xc4\xe7\x7b\xcc\x88\xcc\x20"
"\x45\x9c\x02\xb6\x9b\x4a\xa8\xf5\x82\x17",
.np = 4,
.tap = { 63, 64, 31, 5 }
crypto: testmgr - add test cases for SHA-1, SHA-224, SHA-256 and AES-CCM This adds test cases for SHA-1, SHA-224, SHA-256 and AES-CCM with an input size that is an exact multiple of the block size. The reason is that some implementations use a different code path for these cases. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-16 06:40:04 -06:00
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\xc8\x71\xf6\x9a\x63\xcc\xa9\x84\x84\x82"
"\x64\xe7\x79\x95\x5d\xd7\x19\x41\x7c\x91",
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\xb8\xe3\x54\xed\xc5\xfc\xef\xa4"
"\x55\x73\x4a\x81\x99\xe4\x47\x2a"
"\x30\xd6\xc9\x85",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}
};
/*
* SHA224 test vectors from from FIPS PUB 180-2
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha224_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
.plaintext = "",
.psize = 0,
.digest = "\xd1\x4a\x02\x8c\x2a\x3a\x2b\xc9"
"\x47\x61\x02\xbb\x28\x82\x34\xc4"
"\x15\xa2\xb0\x1f\x82\x8e\xa6\x2a"
"\xc5\xb3\xe4\x2f",
}, {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.plaintext = "abc",
.psize = 3,
.digest = "\x23\x09\x7D\x22\x34\x05\xD8\x22"
"\x86\x42\xA4\x77\xBD\xA2\x55\xB3"
"\x2A\xAD\xBC\xE4\xBD\xA0\xB3\xF7"
"\xE3\x6C\x9D\xA7",
}, {
.plaintext =
"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x75\x38\x8B\x16\x51\x27\x76\xCC"
"\x5D\xBA\x5D\xA1\xFD\x89\x01\x50"
"\xB0\xC6\x45\x5C\xB4\xF5\x8B\x19"
"\x52\x52\x25\x25",
.np = 2,
.tap = { 28, 28 }
crypto: testmgr - add test cases for SHA-1, SHA-224, SHA-256 and AES-CCM This adds test cases for SHA-1, SHA-224, SHA-256 and AES-CCM with an input size that is an exact multiple of the block size. The reason is that some implementations use a different code path for these cases. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-16 06:40:04 -06:00
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\xc4\xdb\x2b\x3a\x58\xc3\x99\x01"
"\x42\xfd\x10\x92\xaa\x4e\x04\x08"
"\x58\xbb\xbb\xe8\xf8\x14\xa7\x0c"
"\xef\x3b\xcb\x0e",
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x98\x43\x07\x63\x75\xe0\xa7\x1c"
"\x78\xb1\x8b\xfd\x04\xf5\x2d\x91"
"\x20\x48\xa4\x28\xff\x55\xb1\xd3"
"\xe6\xf9\x4f\xcc",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}
};
/*
* SHA256 test vectors from from NIST
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha256_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
.plaintext = "",
.psize = 0,
.digest = "\xe3\xb0\xc4\x42\x98\xfc\x1c\x14"
"\x9a\xfb\xf4\xc8\x99\x6f\xb9\x24"
"\x27\xae\x41\xe4\x64\x9b\x93\x4c"
"\xa4\x95\x99\x1b\x78\x52\xb8\x55",
}, {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.plaintext = "abc",
.psize = 3,
.digest = "\xba\x78\x16\xbf\x8f\x01\xcf\xea"
"\x41\x41\x40\xde\x5d\xae\x22\x23"
"\xb0\x03\x61\xa3\x96\x17\x7a\x9c"
"\xb4\x10\xff\x61\xf2\x00\x15\xad",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x24\x8d\x6a\x61\xd2\x06\x38\xb8"
"\xe5\xc0\x26\x93\x0c\x3e\x60\x39"
"\xa3\x3c\xe4\x59\x64\xff\x21\x67"
"\xf6\xec\xed\xd4\x19\xdb\x06\xc1",
.np = 2,
.tap = { 28, 28 }
crypto: testmgr - add test cases for SHA-1, SHA-224, SHA-256 and AES-CCM This adds test cases for SHA-1, SHA-224, SHA-256 and AES-CCM with an input size that is an exact multiple of the block size. The reason is that some implementations use a different code path for these cases. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-16 06:40:04 -06:00
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\xb5\xfe\xad\x56\x7d\xff\xcb\xa4"
"\x2c\x32\x29\x32\x19\xbb\xfb\xfa"
"\xd6\xff\x94\xa3\x72\x91\x85\x66"
"\x3b\xa7\x87\x77\x58\xa3\x40\x3a",
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\xc5\xce\x0c\xca\x01\x4f\x53\x3a"
"\x32\x32\x17\xcc\xd4\x6a\x71\xa9"
"\xf3\xed\x50\x10\x64\x8e\x06\xbe"
"\x9b\x4a\xa6\xbb\x05\x89\x59\x51",
crypto: testmgr - add test cases for SHA-1, SHA-224, SHA-256 and AES-CCM This adds test cases for SHA-1, SHA-224, SHA-256 and AES-CCM with an input size that is an exact multiple of the block size. The reason is that some implementations use a different code path for these cases. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-16 06:40:04 -06:00
}
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
/*
* SHA384 test vectors from from NIST and kerneli
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha384_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
.plaintext = "",
.psize = 0,
.digest = "\x38\xb0\x60\xa7\x51\xac\x96\x38"
"\x4c\xd9\x32\x7e\xb1\xb1\xe3\x6a"
"\x21\xfd\xb7\x11\x14\xbe\x07\x43"
"\x4c\x0c\xc7\xbf\x63\xf6\xe1\xda"
"\x27\x4e\xde\xbf\xe7\x6f\x65\xfb"
"\xd5\x1a\xd2\xf1\x48\x98\xb9\x5b",
}, {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.plaintext= "abc",
.psize = 3,
.digest = "\xcb\x00\x75\x3f\x45\xa3\x5e\x8b"
"\xb5\xa0\x3d\x69\x9a\xc6\x50\x07"
"\x27\x2c\x32\xab\x0e\xde\xd1\x63"
"\x1a\x8b\x60\x5a\x43\xff\x5b\xed"
"\x80\x86\x07\x2b\xa1\xe7\xcc\x23"
"\x58\xba\xec\xa1\x34\xc8\x25\xa7",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x33\x91\xfd\xdd\xfc\x8d\xc7\x39"
"\x37\x07\xa6\x5b\x1b\x47\x09\x39"
"\x7c\xf8\xb1\xd1\x62\xaf\x05\xab"
"\xfe\x8f\x45\x0d\xe5\xf3\x6b\xc6"
"\xb0\x45\x5a\x85\x20\xbc\x4e\x6f"
"\x5f\xe9\x5b\x1f\xe3\xc8\x45\x2b",
}, {
.plaintext = "abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmn"
"hijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu",
.psize = 112,
.digest = "\x09\x33\x0c\x33\xf7\x11\x47\xe8"
"\x3d\x19\x2f\xc7\x82\xcd\x1b\x47"
"\x53\x11\x1b\x17\x3b\x3b\x05\xd2"
"\x2f\xa0\x80\x86\xe3\xb0\xf7\x12"
"\xfc\xc7\xc7\x1a\x55\x7e\x2d\xb9"
"\x66\xc3\xe9\xfa\x91\x74\x60\x39",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcd"
"efghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz",
.psize = 104,
.digest = "\x3d\x20\x89\x73\xab\x35\x08\xdb"
"\xbd\x7e\x2c\x28\x62\xba\x29\x0a"
"\xd3\x01\x0e\x49\x78\xc1\x98\xdc"
"\x4d\x8f\xd0\x14\xe5\x82\x82\x3a"
"\x89\xe1\x6f\x9b\x2a\x7b\xbc\x1a"
"\xc9\x38\xe2\xd1\x99\xe8\xbe\xa4",
.np = 4,
.tap = { 26, 26, 26, 26 }
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x4d\x97\x23\xc8\xea\x7a\x7c\x15"
"\xb8\xff\x97\x9c\xf5\x13\x4f\x31"
"\xde\x67\xf7\x24\x73\xcd\x70\x1c"
"\x03\x4a\xba\x8a\x87\x49\xfe\xdc"
"\x75\x29\x62\x83\xae\x3f\x17\xab"
"\xfd\x10\x4d\x8e\x17\x1c\x1f\xca",
}
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
/*
* SHA512 test vectors from from NIST and kerneli
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec sha512_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
.plaintext = "",
.psize = 0,
.digest = "\xcf\x83\xe1\x35\x7e\xef\xb8\xbd"
"\xf1\x54\x28\x50\xd6\x6d\x80\x07"
"\xd6\x20\xe4\x05\x0b\x57\x15\xdc"
"\x83\xf4\xa9\x21\xd3\x6c\xe9\xce"
"\x47\xd0\xd1\x3c\x5d\x85\xf2\xb0"
"\xff\x83\x18\xd2\x87\x7e\xec\x2f"
"\x63\xb9\x31\xbd\x47\x41\x7a\x81"
"\xa5\x38\x32\x7a\xf9\x27\xda\x3e",
}, {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.plaintext = "abc",
.psize = 3,
.digest = "\xdd\xaf\x35\xa1\x93\x61\x7a\xba"
"\xcc\x41\x73\x49\xae\x20\x41\x31"
"\x12\xe6\xfa\x4e\x89\xa9\x7e\xa2"
"\x0a\x9e\xee\xe6\x4b\x55\xd3\x9a"
"\x21\x92\x99\x2a\x27\x4f\xc1\xa8"
"\x36\xba\x3c\x23\xa3\xfe\xeb\xbd"
"\x45\x4d\x44\x23\x64\x3c\xe8\x0e"
"\x2a\x9a\xc9\x4f\xa5\x4c\xa4\x9f",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x20\x4a\x8f\xc6\xdd\xa8\x2f\x0a"
"\x0c\xed\x7b\xeb\x8e\x08\xa4\x16"
"\x57\xc1\x6e\xf4\x68\xb2\x28\xa8"
"\x27\x9b\xe3\x31\xa7\x03\xc3\x35"
"\x96\xfd\x15\xc1\x3b\x1b\x07\xf9"
"\xaa\x1d\x3b\xea\x57\x78\x9c\xa0"
"\x31\xad\x85\xc7\xa7\x1d\xd7\x03"
"\x54\xec\x63\x12\x38\xca\x34\x45",
}, {
.plaintext = "abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmn"
"hijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu",
.psize = 112,
.digest = "\x8e\x95\x9b\x75\xda\xe3\x13\xda"
"\x8c\xf4\xf7\x28\x14\xfc\x14\x3f"
"\x8f\x77\x79\xc6\xeb\x9f\x7f\xa1"
"\x72\x99\xae\xad\xb6\x88\x90\x18"
"\x50\x1d\x28\x9e\x49\x00\xf7\xe4"
"\x33\x1b\x99\xde\xc4\xb5\x43\x3a"
"\xc7\xd3\x29\xee\xb6\xdd\x26\x54"
"\x5e\x96\xe5\x5b\x87\x4b\xe9\x09",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyzabcd"
"efghijklmnopqrstuvwxyzabcdefghijklmnopqrstuvwxyz",
.psize = 104,
.digest = "\x93\x0d\x0c\xef\xcb\x30\xff\x11"
"\x33\xb6\x89\x81\x21\xf1\xcf\x3d"
"\x27\x57\x8a\xfc\xaf\xe8\x67\x7c"
"\x52\x57\xcf\x06\x99\x11\xf7\x5d"
"\x8f\x58\x31\xb5\x6e\xbf\xda\x67"
"\xb2\x78\xe6\x6d\xff\x8b\x84\xfe"
"\x2b\x28\x70\xf7\x42\xa5\x80\xd8"
"\xed\xb4\x19\x87\x23\x28\x50\xc9",
.np = 4,
.tap = { 26, 26, 26, 26 }
crypto: testmgr - add empty and large test vectors for SHA-1, SHA-224, SHA-256, SHA-384 and SHA-512 Patch adds large test-vectors for SHA algorithms for better code coverage in optimized assembly implementations. Empty test-vectors are also added, as some crypto drivers appear to have special case handling for empty input. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-04-12 06:35:29 -06:00
}, {
.plaintext = "\x08\x9f\x13\xaa\x41\xd8\x4c\xe3"
"\x7a\x11\x85\x1c\xb3\x27\xbe\x55"
"\xec\x60\xf7\x8e\x02\x99\x30\xc7"
"\x3b\xd2\x69\x00\x74\x0b\xa2\x16"
"\xad\x44\xdb\x4f\xe6\x7d\x14\x88"
"\x1f\xb6\x2a\xc1\x58\xef\x63\xfa"
"\x91\x05\x9c\x33\xca\x3e\xd5\x6c"
"\x03\x77\x0e\xa5\x19\xb0\x47\xde"
"\x52\xe9\x80\x17\x8b\x22\xb9\x2d"
"\xc4\x5b\xf2\x66\xfd\x94\x08\x9f"
"\x36\xcd\x41\xd8\x6f\x06\x7a\x11"
"\xa8\x1c\xb3\x4a\xe1\x55\xec\x83"
"\x1a\x8e\x25\xbc\x30\xc7\x5e\xf5"
"\x69\x00\x97\x0b\xa2\x39\xd0\x44"
"\xdb\x72\x09\x7d\x14\xab\x1f\xb6"
"\x4d\xe4\x58\xef\x86\x1d\x91\x28"
"\xbf\x33\xca\x61\xf8\x6c\x03\x9a"
"\x0e\xa5\x3c\xd3\x47\xde\x75\x0c"
"\x80\x17\xae\x22\xb9\x50\xe7\x5b"
"\xf2\x89\x20\x94\x2b\xc2\x36\xcd"
"\x64\xfb\x6f\x06\x9d\x11\xa8\x3f"
"\xd6\x4a\xe1\x78\x0f\x83\x1a\xb1"
"\x25\xbc\x53\xea\x5e\xf5\x8c\x00"
"\x97\x2e\xc5\x39\xd0\x67\xfe\x72"
"\x09\xa0\x14\xab\x42\xd9\x4d\xe4"
"\x7b\x12\x86\x1d\xb4\x28\xbf\x56"
"\xed\x61\xf8\x8f\x03\x9a\x31\xc8"
"\x3c\xd3\x6a\x01\x75\x0c\xa3\x17"
"\xae\x45\xdc\x50\xe7\x7e\x15\x89"
"\x20\xb7\x2b\xc2\x59\xf0\x64\xfb"
"\x92\x06\x9d\x34\xcb\x3f\xd6\x6d"
"\x04\x78\x0f\xa6\x1a\xb1\x48\xdf"
"\x53\xea\x81\x18\x8c\x23\xba\x2e"
"\xc5\x5c\xf3\x67\xfe\x95\x09\xa0"
"\x37\xce\x42\xd9\x70\x07\x7b\x12"
"\xa9\x1d\xb4\x4b\xe2\x56\xed\x84"
"\x1b\x8f\x26\xbd\x31\xc8\x5f\xf6"
"\x6a\x01\x98\x0c\xa3\x3a\xd1\x45"
"\xdc\x73\x0a\x7e\x15\xac\x20\xb7"
"\x4e\xe5\x59\xf0\x87\x1e\x92\x29"
"\xc0\x34\xcb\x62\xf9\x6d\x04\x9b"
"\x0f\xa6\x3d\xd4\x48\xdf\x76\x0d"
"\x81\x18\xaf\x23\xba\x51\xe8\x5c"
"\xf3\x8a\x21\x95\x2c\xc3\x37\xce"
"\x65\xfc\x70\x07\x9e\x12\xa9\x40"
"\xd7\x4b\xe2\x79\x10\x84\x1b\xb2"
"\x26\xbd\x54\xeb\x5f\xf6\x8d\x01"
"\x98\x2f\xc6\x3a\xd1\x68\xff\x73"
"\x0a\xa1\x15\xac\x43\xda\x4e\xe5"
"\x7c\x13\x87\x1e\xb5\x29\xc0\x57"
"\xee\x62\xf9\x90\x04\x9b\x32\xc9"
"\x3d\xd4\x6b\x02\x76\x0d\xa4\x18"
"\xaf\x46\xdd\x51\xe8\x7f\x16\x8a"
"\x21\xb8\x2c\xc3\x5a\xf1\x65\xfc"
"\x93\x07\x9e\x35\xcc\x40\xd7\x6e"
"\x05\x79\x10\xa7\x1b\xb2\x49\xe0"
"\x54\xeb\x82\x19\x8d\x24\xbb\x2f"
"\xc6\x5d\xf4\x68\xff\x96\x0a\xa1"
"\x38\xcf\x43\xda\x71\x08\x7c\x13"
"\xaa\x1e\xb5\x4c\xe3\x57\xee\x85"
"\x1c\x90\x27\xbe\x32\xc9\x60\xf7"
"\x6b\x02\x99\x0d\xa4\x3b\xd2\x46"
"\xdd\x74\x0b\x7f\x16\xad\x21\xb8"
"\x4f\xe6\x5a\xf1\x88\x1f\x93\x2a"
"\xc1\x35\xcc\x63\xfa\x6e\x05\x9c"
"\x10\xa7\x3e\xd5\x49\xe0\x77\x0e"
"\x82\x19\xb0\x24\xbb\x52\xe9\x5d"
"\xf4\x8b\x22\x96\x2d\xc4\x38\xcf"
"\x66\xfd\x71\x08\x9f\x13\xaa\x41"
"\xd8\x4c\xe3\x7a\x11\x85\x1c\xb3"
"\x27\xbe\x55\xec\x60\xf7\x8e\x02"
"\x99\x30\xc7\x3b\xd2\x69\x00\x74"
"\x0b\xa2\x16\xad\x44\xdb\x4f\xe6"
"\x7d\x14\x88\x1f\xb6\x2a\xc1\x58"
"\xef\x63\xfa\x91\x05\x9c\x33\xca"
"\x3e\xd5\x6c\x03\x77\x0e\xa5\x19"
"\xb0\x47\xde\x52\xe9\x80\x17\x8b"
"\x22\xb9\x2d\xc4\x5b\xf2\x66\xfd"
"\x94\x08\x9f\x36\xcd\x41\xd8\x6f"
"\x06\x7a\x11\xa8\x1c\xb3\x4a\xe1"
"\x55\xec\x83\x1a\x8e\x25\xbc\x30"
"\xc7\x5e\xf5\x69\x00\x97\x0b\xa2"
"\x39\xd0\x44\xdb\x72\x09\x7d\x14"
"\xab\x1f\xb6\x4d\xe4\x58\xef\x86"
"\x1d\x91\x28\xbf\x33\xca\x61\xf8"
"\x6c\x03\x9a\x0e\xa5\x3c\xd3\x47"
"\xde\x75\x0c\x80\x17\xae\x22\xb9"
"\x50\xe7\x5b\xf2\x89\x20\x94\x2b"
"\xc2\x36\xcd\x64\xfb\x6f\x06\x9d"
"\x11\xa8\x3f\xd6\x4a\xe1\x78\x0f"
"\x83\x1a\xb1\x25\xbc\x53\xea\x5e"
"\xf5\x8c\x00\x97\x2e\xc5\x39\xd0"
"\x67\xfe\x72\x09\xa0\x14\xab\x42"
"\xd9\x4d\xe4\x7b\x12\x86\x1d\xb4"
"\x28\xbf\x56\xed\x61\xf8\x8f\x03"
"\x9a\x31\xc8\x3c\xd3\x6a\x01\x75"
"\x0c\xa3\x17\xae\x45\xdc\x50\xe7"
"\x7e\x15\x89\x20\xb7\x2b\xc2\x59"
"\xf0\x64\xfb\x92\x06\x9d\x34\xcb"
"\x3f\xd6\x6d\x04\x78\x0f\xa6\x1a"
"\xb1\x48\xdf\x53\xea\x81\x18\x8c"
"\x23\xba\x2e\xc5\x5c\xf3\x67\xfe"
"\x95\x09\xa0\x37\xce\x42\xd9\x70"
"\x07\x7b\x12\xa9\x1d\xb4\x4b\xe2"
"\x56\xed\x84\x1b\x8f\x26\xbd\x31"
"\xc8\x5f\xf6\x6a\x01\x98\x0c\xa3"
"\x3a\xd1\x45\xdc\x73\x0a\x7e\x15"
"\xac\x20\xb7\x4e\xe5\x59\xf0\x87"
"\x1e\x92\x29\xc0\x34\xcb\x62\xf9"
"\x6d\x04\x9b\x0f\xa6\x3d\xd4\x48"
"\xdf\x76\x0d\x81\x18\xaf\x23\xba"
"\x51\xe8\x5c\xf3\x8a\x21\x95\x2c"
"\xc3\x37\xce\x65\xfc\x70\x07\x9e"
"\x12\xa9\x40\xd7\x4b\xe2\x79\x10"
"\x84\x1b\xb2\x26\xbd\x54\xeb\x5f"
"\xf6\x8d\x01\x98\x2f\xc6\x3a\xd1"
"\x68\xff\x73\x0a\xa1\x15\xac\x43"
"\xda\x4e\xe5\x7c\x13\x87\x1e\xb5"
"\x29\xc0\x57\xee\x62\xf9\x90\x04"
"\x9b\x32\xc9\x3d\xd4\x6b\x02\x76"
"\x0d\xa4\x18\xaf\x46\xdd\x51\xe8"
"\x7f\x16\x8a\x21\xb8\x2c\xc3\x5a"
"\xf1\x65\xfc\x93\x07\x9e\x35\xcc"
"\x40\xd7\x6e\x05\x79\x10\xa7\x1b"
"\xb2\x49\xe0\x54\xeb\x82\x19\x8d"
"\x24\xbb\x2f\xc6\x5d\xf4\x68\xff"
"\x96\x0a\xa1\x38\xcf\x43\xda\x71"
"\x08\x7c\x13\xaa\x1e\xb5\x4c",
.psize = 1023,
.digest = "\x76\xc9\xd4\x91\x7a\x5f\x0f\xaa"
"\x13\x39\xf3\x01\x7a\xfa\xe5\x41"
"\x5f\x0b\xf8\xeb\x32\xfc\xbf\xb0"
"\xfa\x8c\xcd\x17\x83\xe2\xfa\xeb"
"\x1c\x19\xde\xe2\x75\xdc\x34\x64"
"\x5f\x35\x9c\x61\x2f\x10\xf9\xec"
"\x59\xca\x9d\xcc\x25\x0c\x43\xba"
"\x85\xa8\xf8\xfe\xb5\x24\xb2\xee",
}
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
/*
* WHIRLPOOL test vectors from Whirlpool package
* by Vincent Rijmen and Paulo S. L. M. Barreto as part of the NESSIE
* submission
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec wp512_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x19\xFA\x61\xD7\x55\x22\xA4\x66"
"\x9B\x44\xE3\x9C\x1D\x2E\x17\x26"
"\xC5\x30\x23\x21\x30\xD4\x07\xF8"
"\x9A\xFE\xE0\x96\x49\x97\xF7\xA7"
"\x3E\x83\xBE\x69\x8B\x28\x8F\xEB"
"\xCF\x88\xE3\xE0\x3C\x4F\x07\x57"
"\xEA\x89\x64\xE5\x9B\x63\xD9\x37"
"\x08\xB1\x38\xCC\x42\xA6\x6E\xB3",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x8A\xCA\x26\x02\x79\x2A\xEC\x6F"
"\x11\xA6\x72\x06\x53\x1F\xB7\xD7"
"\xF0\xDF\xF5\x94\x13\x14\x5E\x69"
"\x73\xC4\x50\x01\xD0\x08\x7B\x42"
"\xD1\x1B\xC6\x45\x41\x3A\xEF\xF6"
"\x3A\x42\x39\x1A\x39\x14\x5A\x59"
"\x1A\x92\x20\x0D\x56\x01\x95\xE5"
"\x3B\x47\x85\x84\xFD\xAE\x23\x1A",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\x4E\x24\x48\xA4\xC6\xF4\x86\xBB"
"\x16\xB6\x56\x2C\x73\xB4\x02\x0B"
"\xF3\x04\x3E\x3A\x73\x1B\xCE\x72"
"\x1A\xE1\xB3\x03\xD9\x7E\x6D\x4C"
"\x71\x81\xEE\xBD\xB6\xC5\x7E\x27"
"\x7D\x0E\x34\x95\x71\x14\xCB\xD6"
"\xC7\x97\xFC\x9D\x95\xD8\xB5\x82"
"\xD2\x25\x29\x20\x76\xD4\xEE\xF5",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x37\x8C\x84\xA4\x12\x6E\x2D\xC6"
"\xE5\x6D\xCC\x74\x58\x37\x7A\xAC"
"\x83\x8D\x00\x03\x22\x30\xF5\x3C"
"\xE1\xF5\x70\x0C\x0F\xFB\x4D\x3B"
"\x84\x21\x55\x76\x59\xEF\x55\xC1"
"\x06\xB4\xB5\x2A\xC5\xA4\xAA\xA6"
"\x92\xED\x92\x00\x52\x83\x8F\x33"
"\x62\xE8\x6D\xBD\x37\xA8\x90\x3E",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xF1\xD7\x54\x66\x26\x36\xFF\xE9"
"\x2C\x82\xEB\xB9\x21\x2A\x48\x4A"
"\x8D\x38\x63\x1E\xAD\x42\x38\xF5"
"\x44\x2E\xE1\x3B\x80\x54\xE4\x1B"
"\x08\xBF\x2A\x92\x51\xC3\x0B\x6A"
"\x0B\x8A\xAE\x86\x17\x7A\xB4\xA6"
"\xF6\x8F\x67\x3E\x72\x07\x86\x5D"
"\x5D\x98\x19\xA3\xDB\xA4\xEB\x3B",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
"abcdefghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xDC\x37\xE0\x08\xCF\x9E\xE6\x9B"
"\xF1\x1F\x00\xED\x9A\xBA\x26\x90"
"\x1D\xD7\xC2\x8C\xDE\xC0\x66\xCC"
"\x6A\xF4\x2E\x40\xF8\x2F\x3A\x1E"
"\x08\xEB\xA2\x66\x29\x12\x9D\x8F"
"\xB7\xCB\x57\x21\x1B\x92\x81\xA6"
"\x55\x17\xCC\x87\x9D\x7B\x96\x21"
"\x42\xC6\x5F\x5A\x7A\xF0\x14\x67",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x46\x6E\xF1\x8B\xAB\xB0\x15\x4D"
"\x25\xB9\xD3\x8A\x64\x14\xF5\xC0"
"\x87\x84\x37\x2B\xCC\xB2\x04\xD6"
"\x54\x9C\x4A\xFA\xDB\x60\x14\x29"
"\x4D\x5B\xD8\xDF\x2A\x6C\x44\xE5"
"\x38\xCD\x04\x7B\x26\x81\xA5\x1A"
"\x2C\x60\x48\x1E\x88\xC5\xA2\x0B"
"\x2C\x2A\x80\xCF\x3A\x9A\x08\x3B",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijk",
.psize = 32,
.digest = "\x2A\x98\x7E\xA4\x0F\x91\x70\x61"
"\xF5\xD6\xF0\xA0\xE4\x64\x4F\x48"
"\x8A\x7A\x5A\x52\xDE\xEE\x65\x62"
"\x07\xC5\x62\xF9\x88\xE9\x5C\x69"
"\x16\xBD\xC8\x03\x1B\xC5\xBE\x1B"
"\x7B\x94\x76\x39\xFE\x05\x0B\x56"
"\x93\x9B\xAA\xA0\xAD\xFF\x9A\xE6"
"\x74\x5B\x7B\x18\x1C\x3B\xE3\xFD",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec wp384_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x19\xFA\x61\xD7\x55\x22\xA4\x66"
"\x9B\x44\xE3\x9C\x1D\x2E\x17\x26"
"\xC5\x30\x23\x21\x30\xD4\x07\xF8"
"\x9A\xFE\xE0\x96\x49\x97\xF7\xA7"
"\x3E\x83\xBE\x69\x8B\x28\x8F\xEB"
"\xCF\x88\xE3\xE0\x3C\x4F\x07\x57",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x8A\xCA\x26\x02\x79\x2A\xEC\x6F"
"\x11\xA6\x72\x06\x53\x1F\xB7\xD7"
"\xF0\xDF\xF5\x94\x13\x14\x5E\x69"
"\x73\xC4\x50\x01\xD0\x08\x7B\x42"
"\xD1\x1B\xC6\x45\x41\x3A\xEF\xF6"
"\x3A\x42\x39\x1A\x39\x14\x5A\x59",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\x4E\x24\x48\xA4\xC6\xF4\x86\xBB"
"\x16\xB6\x56\x2C\x73\xB4\x02\x0B"
"\xF3\x04\x3E\x3A\x73\x1B\xCE\x72"
"\x1A\xE1\xB3\x03\xD9\x7E\x6D\x4C"
"\x71\x81\xEE\xBD\xB6\xC5\x7E\x27"
"\x7D\x0E\x34\x95\x71\x14\xCB\xD6",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x37\x8C\x84\xA4\x12\x6E\x2D\xC6"
"\xE5\x6D\xCC\x74\x58\x37\x7A\xAC"
"\x83\x8D\x00\x03\x22\x30\xF5\x3C"
"\xE1\xF5\x70\x0C\x0F\xFB\x4D\x3B"
"\x84\x21\x55\x76\x59\xEF\x55\xC1"
"\x06\xB4\xB5\x2A\xC5\xA4\xAA\xA6",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xF1\xD7\x54\x66\x26\x36\xFF\xE9"
"\x2C\x82\xEB\xB9\x21\x2A\x48\x4A"
"\x8D\x38\x63\x1E\xAD\x42\x38\xF5"
"\x44\x2E\xE1\x3B\x80\x54\xE4\x1B"
"\x08\xBF\x2A\x92\x51\xC3\x0B\x6A"
"\x0B\x8A\xAE\x86\x17\x7A\xB4\xA6",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
"abcdefghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xDC\x37\xE0\x08\xCF\x9E\xE6\x9B"
"\xF1\x1F\x00\xED\x9A\xBA\x26\x90"
"\x1D\xD7\xC2\x8C\xDE\xC0\x66\xCC"
"\x6A\xF4\x2E\x40\xF8\x2F\x3A\x1E"
"\x08\xEB\xA2\x66\x29\x12\x9D\x8F"
"\xB7\xCB\x57\x21\x1B\x92\x81\xA6",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x46\x6E\xF1\x8B\xAB\xB0\x15\x4D"
"\x25\xB9\xD3\x8A\x64\x14\xF5\xC0"
"\x87\x84\x37\x2B\xCC\xB2\x04\xD6"
"\x54\x9C\x4A\xFA\xDB\x60\x14\x29"
"\x4D\x5B\xD8\xDF\x2A\x6C\x44\xE5"
"\x38\xCD\x04\x7B\x26\x81\xA5\x1A",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijk",
.psize = 32,
.digest = "\x2A\x98\x7E\xA4\x0F\x91\x70\x61"
"\xF5\xD6\xF0\xA0\xE4\x64\x4F\x48"
"\x8A\x7A\x5A\x52\xDE\xEE\x65\x62"
"\x07\xC5\x62\xF9\x88\xE9\x5C\x69"
"\x16\xBD\xC8\x03\x1B\xC5\xBE\x1B"
"\x7B\x94\x76\x39\xFE\x05\x0B\x56",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec wp256_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x19\xFA\x61\xD7\x55\x22\xA4\x66"
"\x9B\x44\xE3\x9C\x1D\x2E\x17\x26"
"\xC5\x30\x23\x21\x30\xD4\x07\xF8"
"\x9A\xFE\xE0\x96\x49\x97\xF7\xA7",
}, {
.plaintext = "a",
.psize = 1,
.digest = "\x8A\xCA\x26\x02\x79\x2A\xEC\x6F"
"\x11\xA6\x72\x06\x53\x1F\xB7\xD7"
"\xF0\xDF\xF5\x94\x13\x14\x5E\x69"
"\x73\xC4\x50\x01\xD0\x08\x7B\x42",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\x4E\x24\x48\xA4\xC6\xF4\x86\xBB"
"\x16\xB6\x56\x2C\x73\xB4\x02\x0B"
"\xF3\x04\x3E\x3A\x73\x1B\xCE\x72"
"\x1A\xE1\xB3\x03\xD9\x7E\x6D\x4C",
}, {
.plaintext = "message digest",
.psize = 14,
.digest = "\x37\x8C\x84\xA4\x12\x6E\x2D\xC6"
"\xE5\x6D\xCC\x74\x58\x37\x7A\xAC"
"\x83\x8D\x00\x03\x22\x30\xF5\x3C"
"\xE1\xF5\x70\x0C\x0F\xFB\x4D\x3B",
}, {
.plaintext = "abcdefghijklmnopqrstuvwxyz",
.psize = 26,
.digest = "\xF1\xD7\x54\x66\x26\x36\xFF\xE9"
"\x2C\x82\xEB\xB9\x21\x2A\x48\x4A"
"\x8D\x38\x63\x1E\xAD\x42\x38\xF5"
"\x44\x2E\xE1\x3B\x80\x54\xE4\x1B",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
"abcdefghijklmnopqrstuvwxyz0123456789",
.psize = 62,
.digest = "\xDC\x37\xE0\x08\xCF\x9E\xE6\x9B"
"\xF1\x1F\x00\xED\x9A\xBA\x26\x90"
"\x1D\xD7\xC2\x8C\xDE\xC0\x66\xCC"
"\x6A\xF4\x2E\x40\xF8\x2F\x3A\x1E",
}, {
.plaintext = "1234567890123456789012345678901234567890"
"1234567890123456789012345678901234567890",
.psize = 80,
.digest = "\x46\x6E\xF1\x8B\xAB\xB0\x15\x4D"
"\x25\xB9\xD3\x8A\x64\x14\xF5\xC0"
"\x87\x84\x37\x2B\xCC\xB2\x04\xD6"
"\x54\x9C\x4A\xFA\xDB\x60\x14\x29",
}, {
.plaintext = "abcdbcdecdefdefgefghfghighijhijk",
.psize = 32,
.digest = "\x2A\x98\x7E\xA4\x0F\x91\x70\x61"
"\xF5\xD6\xF0\xA0\xE4\x64\x4F\x48"
"\x8A\x7A\x5A\x52\xDE\xEE\x65\x62"
"\x07\xC5\x62\xF9\x88\xE9\x5C\x69",
},
};
/*
* TIGER test vectors from Tiger website
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec tgr192_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x24\xf0\x13\x0c\x63\xac\x93\x32"
"\x16\x16\x6e\x76\xb1\xbb\x92\x5f"
"\xf3\x73\xde\x2d\x49\x58\x4e\x7a",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xf2\x58\xc1\xe8\x84\x14\xab\x2a"
"\x52\x7a\xb5\x41\xff\xc5\xb8\xbf"
"\x93\x5f\x7b\x95\x1c\x13\x29\x51",
}, {
.plaintext = "Tiger",
.psize = 5,
.digest = "\x9f\x00\xf5\x99\x07\x23\x00\xdd"
"\x27\x6a\xbb\x38\xc8\xeb\x6d\xec"
"\x37\x79\x0c\x11\x6f\x9d\x2b\xdf",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\x87\xfb\x2a\x90\x83\x85\x1c\xf7"
"\x47\x0d\x2c\xf8\x10\xe6\xdf\x9e"
"\xb5\x86\x44\x50\x34\xa5\xa3\x86",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ=abcdefghijklmnopqrstuvwxyz+0123456789",
.psize = 64,
.digest = "\x46\x7d\xb8\x08\x63\xeb\xce\x48"
"\x8d\xf1\xcd\x12\x61\x65\x5d\xe9"
"\x57\x89\x65\x65\x97\x5f\x91\x97",
}, {
.plaintext = "Tiger - A Fast New Hash Function, "
"by Ross Anderson and Eli Biham, "
"proceedings of Fast Software Encryption 3, "
"Cambridge, 1996.",
.psize = 125,
.digest = "\x3d\x9a\xeb\x03\xd1\xbd\x1a\x63"
"\x57\xb2\x77\x4d\xfd\x6d\x5b\x24"
"\xdd\x68\x15\x1d\x50\x39\x74\xfc",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec tgr160_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x24\xf0\x13\x0c\x63\xac\x93\x32"
"\x16\x16\x6e\x76\xb1\xbb\x92\x5f"
"\xf3\x73\xde\x2d",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xf2\x58\xc1\xe8\x84\x14\xab\x2a"
"\x52\x7a\xb5\x41\xff\xc5\xb8\xbf"
"\x93\x5f\x7b\x95",
}, {
.plaintext = "Tiger",
.psize = 5,
.digest = "\x9f\x00\xf5\x99\x07\x23\x00\xdd"
"\x27\x6a\xbb\x38\xc8\xeb\x6d\xec"
"\x37\x79\x0c\x11",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\x87\xfb\x2a\x90\x83\x85\x1c\xf7"
"\x47\x0d\x2c\xf8\x10\xe6\xdf\x9e"
"\xb5\x86\x44\x50",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ=abcdefghijklmnopqrstuvwxyz+0123456789",
.psize = 64,
.digest = "\x46\x7d\xb8\x08\x63\xeb\xce\x48"
"\x8d\xf1\xcd\x12\x61\x65\x5d\xe9"
"\x57\x89\x65\x65",
}, {
.plaintext = "Tiger - A Fast New Hash Function, "
"by Ross Anderson and Eli Biham, "
"proceedings of Fast Software Encryption 3, "
"Cambridge, 1996.",
.psize = 125,
.digest = "\x3d\x9a\xeb\x03\xd1\xbd\x1a\x63"
"\x57\xb2\x77\x4d\xfd\x6d\x5b\x24"
"\xdd\x68\x15\x1d",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec tgr128_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.plaintext = "",
.psize = 0,
.digest = "\x24\xf0\x13\x0c\x63\xac\x93\x32"
"\x16\x16\x6e\x76\xb1\xbb\x92\x5f",
}, {
.plaintext = "abc",
.psize = 3,
.digest = "\xf2\x58\xc1\xe8\x84\x14\xab\x2a"
"\x52\x7a\xb5\x41\xff\xc5\xb8\xbf",
}, {
.plaintext = "Tiger",
.psize = 5,
.digest = "\x9f\x00\xf5\x99\x07\x23\x00\xdd"
"\x27\x6a\xbb\x38\xc8\xeb\x6d\xec",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+-",
.psize = 64,
.digest = "\x87\xfb\x2a\x90\x83\x85\x1c\xf7"
"\x47\x0d\x2c\xf8\x10\xe6\xdf\x9e",
}, {
.plaintext = "ABCDEFGHIJKLMNOPQRSTUVWXYZ=abcdefghijklmnopqrstuvwxyz+0123456789",
.psize = 64,
.digest = "\x46\x7d\xb8\x08\x63\xeb\xce\x48"
"\x8d\xf1\xcd\x12\x61\x65\x5d\xe9",
}, {
.plaintext = "Tiger - A Fast New Hash Function, "
"by Ross Anderson and Eli Biham, "
"proceedings of Fast Software Encryption 3, "
"Cambridge, 1996.",
.psize = 125,
.digest = "\x3d\x9a\xeb\x03\xd1\xbd\x1a\x63"
"\x57\xb2\x77\x4d\xfd\x6d\x5b\x24",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec ghash_tv_template[] =
crypto: testmgr - Add ghash algorithm test before provide to users Add ghash algorithm test before provide it to users Signed-off-by: Youquan, Song <youquan.song@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-11-23 05:23:04 -07:00
{
{
crypto: testmgr - add 4 more test vectors for GHASH This adds 4 test vectors for GHASH (of which one for chunked mode), making a total of 5. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-06-12 09:01:50 -06:00
.key = "\xdf\xa6\xbf\x4d\xed\x81\xdb\x03"
"\xff\xca\xff\x95\xf8\x30\xf0\x61",
crypto: testmgr - Add ghash algorithm test before provide to users Add ghash algorithm test before provide it to users Signed-off-by: Youquan, Song <youquan.song@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-11-23 05:23:04 -07:00
.ksize = 16,
crypto: testmgr - add 4 more test vectors for GHASH This adds 4 test vectors for GHASH (of which one for chunked mode), making a total of 5. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-06-12 09:01:50 -06:00
.plaintext = "\x95\x2b\x2a\x56\xa5\x60\x04a\xc0"
"\xb3\x2b\x66\x56\xa0\x5b\x40\xb6",
crypto: testmgr - Add ghash algorithm test before provide to users Add ghash algorithm test before provide it to users Signed-off-by: Youquan, Song <youquan.song@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-11-23 05:23:04 -07:00
.psize = 16,
.digest = "\xda\x53\xeb\x0a\xd2\xc5\x5b\xb6"
"\x4f\xc4\x80\x2c\xc3\xfe\xda\x60",
crypto: testmgr - add 4 more test vectors for GHASH This adds 4 test vectors for GHASH (of which one for chunked mode), making a total of 5. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-06-12 09:01:50 -06:00
}, {
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 16,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x3e\x1f\x5c\x4d\x65\xf0\xef\xce"
"\x0d\x61\x06\x27\x66\x51\xd5\xe2",
.np = 2,
.tap = {14, 14}
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 16,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\xfb\x49\x8a\x36\xe1\x96\xe1\x96"
"\xe1\x96\xe1\x96\xe1\x96\xe1\x96",
}, {
.key = "\xda\x53\xeb\x0a\xd2\xc5\x5b\xb6"
"\x4f\xc4\x80\x2c\xc3\xfe\xda\x60",
.ksize = 16,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\x2b\x5c\x0c\x7f\x52\xd1\x60\xc2"
"\x49\xed\x6e\x32\x7a\xa9\xbe\x08",
}, {
.key = "\x95\x2b\x2a\x56\xa5\x60\x04a\xc0"
"\xb3\x2b\x66\x56\xa0\x5b\x40\xb6",
.ksize = 16,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\xf8\x94\x87\x2a\x4b\x63\x99\x28"
"\x23\xf7\x93\xf7\x19\xf5\x96\xd9",
crypto: testmgr - Added one larger ghash testvector (400 bytes) to the testmgr. Signed-off-by: Harald Freudenberger <freude@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-05-21 09:34:31 -06:00
}, {
.key = "\x0a\x1b\x2c\x3d\x4e\x5f\x64\x71"
"\x82\x93\xa4\xb5\xc6\xd7\xe8\xf9",
.ksize = 16,
.plaintext = "\x56\x6f\x72\x20\x6c\x61\x75\x74"
"\x65\x72\x20\x4c\x61\x75\x73\x63"
"\x68\x65\x6e\x20\x75\x6e\x64\x20"
"\x53\x74\x61\x75\x6e\x65\x6e\x20"
"\x73\x65\x69\x20\x73\x74\x69\x6c"
"\x6c\x2c\x0a\x64\x75\x20\x6d\x65"
"\x69\x6e\x20\x74\x69\x65\x66\x74"
"\x69\x65\x66\x65\x73\x20\x4c\x65"
"\x62\x65\x6e\x3b\x0a\x64\x61\x73"
"\x73\x20\x64\x75\x20\x77\x65\x69"
"\xc3\x9f\x74\x20\x77\x61\x73\x20"
"\x64\x65\x72\x20\x57\x69\x6e\x64"
"\x20\x64\x69\x72\x20\x77\x69\x6c"
"\x6c\x2c\x0a\x65\x68\x20\x6e\x6f"
"\x63\x68\x20\x64\x69\x65\x20\x42"
"\x69\x72\x6b\x65\x6e\x20\x62\x65"
"\x62\x65\x6e\x2e\x0a\x0a\x55\x6e"
"\x64\x20\x77\x65\x6e\x6e\x20\x64"
"\x69\x72\x20\x65\x69\x6e\x6d\x61"
"\x6c\x20\x64\x61\x73\x20\x53\x63"
"\x68\x77\x65\x69\x67\x65\x6e\x20"
"\x73\x70\x72\x61\x63\x68\x2c\x0a"
"\x6c\x61\x73\x73\x20\x64\x65\x69"
"\x6e\x65\x20\x53\x69\x6e\x6e\x65"
"\x20\x62\x65\x73\x69\x65\x67\x65"
"\x6e\x2e\x0a\x4a\x65\x64\x65\x6d"
"\x20\x48\x61\x75\x63\x68\x65\x20"
"\x67\x69\x62\x74\x20\x64\x69\x63"
"\x68\x2c\x20\x67\x69\x62\x20\x6e"
"\x61\x63\x68\x2c\x0a\x65\x72\x20"
"\x77\x69\x72\x64\x20\x64\x69\x63"
"\x68\x20\x6c\x69\x65\x62\x65\x6e"
"\x20\x75\x6e\x64\x20\x77\x69\x65"
"\x67\x65\x6e\x2e\x0a\x0a\x55\x6e"
"\x64\x20\x64\x61\x6e\x6e\x20\x6d"
"\x65\x69\x6e\x65\x20\x53\x65\x65"
"\x6c\x65\x20\x73\x65\x69\x74\x20"
"\x77\x65\x69\x74\x2c\x20\x73\x65"
"\x69\x20\x77\x65\x69\x74\x2c\x0a"
"\x64\x61\x73\x73\x20\x64\x69\x72"
"\x20\x64\x61\x73\x20\x4c\x65\x62"
"\x65\x6e\x20\x67\x65\x6c\x69\x6e"
"\x67\x65\x2c\x0a\x62\x72\x65\x69"
"\x74\x65\x20\x64\x69\x63\x68\x20"
"\x77\x69\x65\x20\x65\x69\x6e\x20"
"\x46\x65\x69\x65\x72\x6b\x6c\x65"
"\x69\x64\x0a\xc3\xbc\x62\x65\x72"
"\x20\x64\x69\x65\x20\x73\x69\x6e"
"\x6e\x65\x6e\x64\x65\x6e\x20\x44"
"\x69\x6e\x67\x65\x2e\x2e\x2e\x0a",
.psize = 400,
.digest = "\xad\xb1\xc1\xe9\x56\x70\x31\x1d"
"\xbb\x5b\xdf\x5e\x70\x72\x1a\x57",
crypto: testmgr - Add ghash algorithm test before provide to users Add ghash algorithm test before provide it to users Signed-off-by: Youquan, Song <youquan.song@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-11-23 05:23:04 -07:00
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* HMAC-MD5 test vectors from RFC2202
* (These need to be fixed to not use strlen).
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_md5_tv_template[] =
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 16,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x92\x94\x72\x7a\x36\x38\xbb\x1c"
"\x13\xf4\x8e\xf8\x15\x8b\xfc\x9d",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x75\x0c\x78\x3e\x6a\xb0\xb5\x03"
"\xea\xa8\x6e\x31\x0a\x5d\xb7\x38",
.np = 2,
.tap = {14, 14}
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 16,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\x56\xbe\x34\x52\x1d\x14\x4c\x88"
"\xdb\xb8\xc7\x33\xf0\xe8\xb3\xf6",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18\x19",
.ksize = 25,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\x69\x7e\xaf\x0a\xca\x3a\x3a\xea"
"\x3a\x75\x16\x47\x46\xff\xaa\x79",
}, {
.key = "\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c",
.ksize = 16,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\x56\x46\x1e\xf2\x34\x2e\xdc\x00"
"\xf9\xba\xb9\x95\x69\x0e\xfd\x4c",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key - Hash Key First",
.psize = 54,
.digest = "\x6b\x1a\xb7\xfe\x4b\xd7\xbf\x8f"
"\x0b\x62\xe6\xce\x61\xb9\xd0\xcd",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key and Larger Than One "
"Block-Size Data",
.psize = 73,
.digest = "\x6f\x63\x0f\xad\x67\xcd\xa0\xee"
"\x1f\xb1\xf5\x62\xdb\x3a\xa5\x3e",
},
};
/*
* HMAC-RIPEMD128 test vectors from RFC2286
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_rmd128_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 16,
.plaintext = "Hi There",
.psize = 8,
.digest = "\xfb\xf6\x1f\x94\x92\xaa\x4b\xbf"
"\x81\xc1\x72\xe8\x4e\x07\x34\xdb",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x87\x5f\x82\x88\x62\xb6\xb3\x34"
"\xb4\x27\xc5\x5f\x9f\x7f\xf0\x9b",
.np = 2,
.tap = { 14, 14 },
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 16,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\x09\xf0\xb2\x84\x6d\x2f\x54\x3d"
"\xa3\x63\xcb\xec\x8d\x62\xa3\x8d",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18\x19",
.ksize = 25,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\xbd\xbb\xd7\xcf\x03\xe4\x4b\x5a"
"\xa6\x0a\xf8\x15\xbe\x4d\x22\x94",
}, {
.key = "\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c",
.ksize = 16,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\xe7\x98\x08\xf2\x4b\x25\xfd\x03"
"\x1c\x15\x5f\x0d\x55\x1d\x9a\x3a",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key - Hash Key First",
.psize = 54,
.digest = "\xdc\x73\x29\x28\xde\x98\x10\x4a"
"\x1f\x59\xd3\x73\xc1\x50\xac\xbb",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key and Larger Than One "
"Block-Size Data",
.psize = 73,
.digest = "\x5c\x6b\xec\x96\x79\x3e\x16\xd4"
"\x06\x90\xc2\x37\x63\x5f\x30\xc5",
},
};
/*
* HMAC-RIPEMD160 test vectors from RFC2286
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_rmd160_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x24\xcb\x4b\xd6\x7d\x20\xfc\x1a\x5d\x2e"
"\xd7\x73\x2d\xcc\x39\x37\x7f\x0a\x56\x68",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\xdd\xa6\xc0\x21\x3a\x48\x5a\x9e\x24\xf4"
"\x74\x20\x64\xa7\xf0\x33\xb4\x3c\x40\x69",
.np = 2,
.tap = { 14, 14 },
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 20,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\xb0\xb1\x05\x36\x0d\xe7\x59\x96\x0a\xb4"
"\xf3\x52\x98\xe1\x16\xe2\x95\xd8\xe7\xc1",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18\x19",
.ksize = 25,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\xd5\xca\x86\x2f\x4d\x21\xd5\xe6\x10\xe1"
"\x8b\x4c\xf1\xbe\xb9\x7a\x43\x65\xec\xf4",
}, {
.key = "\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c",
.ksize = 20,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\x76\x19\x69\x39\x78\xf9\x1d\x90\x53\x9a"
"\xe7\x86\x50\x0f\xf3\xd8\xe0\x51\x8e\x39",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key - Hash Key First",
.psize = 54,
.digest = "\x64\x66\xca\x07\xac\x5e\xac\x29\xe1\xbd"
"\x52\x3e\x5a\xda\x76\x05\xb7\x91\xfd\x8b",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key and Larger Than One "
"Block-Size Data",
.psize = 73,
.digest = "\x69\xea\x60\x79\x8d\x71\x61\x6c\xce\x5f"
"\xd0\x87\x1e\x23\x75\x4c\xd7\x5d\x5a\x0a",
},
};
/*
* HMAC-SHA1 test vectors from RFC2202
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha1_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\xb6\x17\x31\x86\x55\x05\x72\x64"
"\xe2\x8b\xc0\xb6\xfb\x37\x8c\x8e\xf1"
"\x46\xbe",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\xef\xfc\xdf\x6a\xe5\xeb\x2f\xa2\xd2\x74"
"\x16\xd5\xf1\x84\xdf\x9c\x25\x9a\x7c\x79",
.np = 2,
.tap = { 14, 14 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 20,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\x12\x5d\x73\x42\xb9\xac\x11\xcd\x91\xa3"
"\x9a\xf4\x8a\xa1\x7b\x4f\x63\xf1\x75\xd3",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18\x19",
.ksize = 25,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\x4c\x90\x07\xf4\x02\x62\x50\xc6\xbc\x84"
"\x14\xf9\xbf\x50\xc8\x6c\x2d\x72\x35\xda",
}, {
.key = "\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c",
.ksize = 20,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\x4c\x1a\x03\x42\x4b\x55\xe0\x7f\xe7\xf2"
"\x7b\xe1\xd5\x8b\xb9\x32\x4a\x9a\x5a\x04",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key - Hash Key First",
.psize = 54,
.digest = "\xaa\x4a\xe5\xe1\x52\x72\xd0\x0e\x95\x70"
"\x56\x37\xce\x8a\x3b\x55\xed\x40\x21\x12",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key and Larger Than One "
"Block-Size Data",
.psize = 73,
.digest = "\xe8\xe9\x9d\x0f\x45\x23\x7d\x78\x6d\x6b"
"\xba\xa7\x96\x5c\x78\x08\xbb\xff\x1a\x91",
},
};
/*
* SHA224 HMAC test vectors from RFC4231
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha224_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
/* ("Hi There") */
.plaintext = "\x48\x69\x20\x54\x68\x65\x72\x65",
.psize = 8,
.digest = "\x89\x6f\xb1\x12\x8a\xbb\xdf\x19"
"\x68\x32\x10\x7c\xd4\x9d\xf3\x3f"
"\x47\xb4\xb1\x16\x99\x12\xba\x4f"
"\x53\x68\x4b\x22",
}, {
.key = "Jefe",
.ksize = 4,
/* ("what do ya want for nothing?") */
.plaintext = "\x77\x68\x61\x74\x20\x64\x6f\x20"
"\x79\x61\x20\x77\x61\x6e\x74\x20"
"\x66\x6f\x72\x20\x6e\x6f\x74\x68"
"\x69\x6e\x67\x3f",
.psize = 28,
.digest = "\xa3\x0e\x01\x09\x8b\xc6\xdb\xbf"
"\x45\x69\x0f\x3a\x7e\x9e\x6d\x0f"
"\x8b\xbe\xa2\xa3\x9e\x61\x48\x00"
"\x8f\xd0\x5e\x44",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
/* ("Test Using Larger Than Block-Size Key - Hash Key First") */
.plaintext = "\x54\x65\x73\x74\x20\x55\x73\x69"
"\x6e\x67\x20\x4c\x61\x72\x67\x65"
"\x72\x20\x54\x68\x61\x6e\x20\x42"
"\x6c\x6f\x63\x6b\x2d\x53\x69\x7a"
"\x65\x20\x4b\x65\x79\x20\x2d\x20"
"\x48\x61\x73\x68\x20\x4b\x65\x79"
"\x20\x46\x69\x72\x73\x74",
.psize = 54,
.digest = "\x95\xe9\xa0\xdb\x96\x20\x95\xad"
"\xae\xbe\x9b\x2d\x6f\x0d\xbc\xe2"
"\xd4\x99\xf1\x12\xf2\xd2\xb7\x27"
"\x3f\xa6\x87\x0e",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
/* ("This is a test using a larger than block-size key and a")
(" larger than block-size data. The key needs to be")
(" hashed before being used by the HMAC algorithm.") */
.plaintext = "\x54\x68\x69\x73\x20\x69\x73\x20"
"\x61\x20\x74\x65\x73\x74\x20\x75"
"\x73\x69\x6e\x67\x20\x61\x20\x6c"
"\x61\x72\x67\x65\x72\x20\x74\x68"
"\x61\x6e\x20\x62\x6c\x6f\x63\x6b"
"\x2d\x73\x69\x7a\x65\x20\x6b\x65"
"\x79\x20\x61\x6e\x64\x20\x61\x20"
"\x6c\x61\x72\x67\x65\x72\x20\x74"
"\x68\x61\x6e\x20\x62\x6c\x6f\x63"
"\x6b\x2d\x73\x69\x7a\x65\x20\x64"
"\x61\x74\x61\x2e\x20\x54\x68\x65"
"\x20\x6b\x65\x79\x20\x6e\x65\x65"
"\x64\x73\x20\x74\x6f\x20\x62\x65"
"\x20\x68\x61\x73\x68\x65\x64\x20"
"\x62\x65\x66\x6f\x72\x65\x20\x62"
"\x65\x69\x6e\x67\x20\x75\x73\x65"
"\x64\x20\x62\x79\x20\x74\x68\x65"
"\x20\x48\x4d\x41\x43\x20\x61\x6c"
"\x67\x6f\x72\x69\x74\x68\x6d\x2e",
.psize = 152,
.digest = "\x3a\x85\x41\x66\xac\x5d\x9f\x02"
"\x3f\x54\xd5\x17\xd0\xb3\x9d\xbd"
"\x94\x67\x70\xdb\x9c\x2b\x95\xc9"
"\xf6\xf5\x65\xd1",
},
};
/*
* HMAC-SHA256 test vectors from
* draft-ietf-ipsec-ciph-sha-256-01.txt
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha256_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20",
.ksize = 32,
.plaintext = "abc",
.psize = 3,
.digest = "\xa2\x1b\x1f\x5d\x4c\xf4\xf7\x3a"
"\x4d\xd9\x39\x75\x0f\x7a\x06\x6a"
"\x7f\x98\xcc\x13\x1c\xb1\x6a\x66"
"\x92\x75\x90\x21\xcf\xab\x81\x81",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20",
.ksize = 32,
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 56,
.digest = "\x10\x4f\xdc\x12\x57\x32\x8f\x08"
"\x18\x4b\xa7\x31\x31\xc5\x3c\xae"
"\xe6\x98\xe3\x61\x19\x42\x11\x49"
"\xea\x8c\x71\x24\x56\x69\x7d\x30",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20",
.ksize = 32,
.plaintext = "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq"
"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
.psize = 112,
.digest = "\x47\x03\x05\xfc\x7e\x40\xfe\x34"
"\xd3\xee\xb3\xe7\x73\xd9\x5a\xab"
"\x73\xac\xf0\xfd\x06\x04\x47\xa5"
"\xeb\x45\x95\xbf\x33\xa9\xd1\xa3",
}, {
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b",
.ksize = 32,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x19\x8a\x60\x7e\xb4\x4b\xfb\xc6"
"\x99\x03\xa0\xf1\xcf\x2b\xbd\xc5"
"\xba\x0a\xa3\xf3\xd9\xae\x3c\x1c"
"\x7a\x3b\x16\x96\xa0\xb6\x8c\xf7",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x5b\xdc\xc1\x46\xbf\x60\x75\x4e"
"\x6a\x04\x24\x26\x08\x95\x75\xc7"
"\x5a\x00\x3f\x08\x9d\x27\x39\x83"
"\x9d\xec\x58\xb9\x64\xec\x38\x43",
.np = 2,
.tap = { 14, 14 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa",
.ksize = 32,
.plaintext = "\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd"
"\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd\xdd",
.psize = 50,
.digest = "\xcd\xcb\x12\x20\xd1\xec\xcc\xea"
"\x91\xe5\x3a\xba\x30\x92\xf9\x62"
"\xe5\x49\xfe\x6c\xe9\xed\x7f\xdc"
"\x43\x19\x1f\xbd\xe4\x5c\x30\xb0",
}, {
.key = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
"\x21\x22\x23\x24\x25",
.ksize = 37,
.plaintext = "\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd"
"\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd\xcd",
.psize = 50,
.digest = "\xd4\x63\x3c\x17\xf6\xfb\x8d\x74"
"\x4c\x66\xde\xe0\xf8\xf0\x74\x55"
"\x6e\xc4\xaf\x55\xef\x07\x99\x85"
"\x41\x46\x8e\xb4\x9b\xd2\xe9\x17",
}, {
.key = "\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c"
"\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c\x0c"
"\x0c\x0c\x0c\x0c\x0c\x0c",
.ksize = 32,
.plaintext = "Test With Truncation",
.psize = 20,
.digest = "\x75\x46\xaf\x01\x84\x1f\xc0\x9b"
"\x1a\xb9\xc3\x74\x9a\x5f\x1c\x17"
"\xd4\xf5\x89\x66\x8a\x58\x7b\x27"
"\x00\xa9\xc9\x7c\x11\x93\xcf\x42",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key - Hash Key First",
.psize = 54,
.digest = "\x69\x53\x02\x5e\xd9\x6f\x0c\x09"
"\xf8\x0a\x96\xf7\x8e\x65\x38\xdb"
"\xe2\xe7\xb8\x20\xe3\xdd\x97\x0e"
"\x7d\xdd\x39\x09\x1b\x32\x35\x2f",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa",
.ksize = 80,
.plaintext = "Test Using Larger Than Block-Size Key and Larger Than "
"One Block-Size Data",
.psize = 73,
.digest = "\x63\x55\xac\x22\xe8\x90\xd0\xa3"
"\xc8\x48\x1a\x5c\xa4\x82\x5b\xc8"
"\x84\xd3\xe7\xa1\xff\x98\xa2\xfc"
"\x2a\xc7\xd8\xe0\x64\xc3\xb2\xe6",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec aes_cmac128_tv_template[] = {
crypto: add CMAC support to CryptoAPI Patch adds support for NIST recommended block cipher mode CMAC to CryptoAPI. This work is based on Tom St Denis' earlier patch, http://marc.info/?l=linux-crypto-vger&m=135877306305466&w=2 Cc: Tom St Denis <tstdenis@elliptictech.com> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-08 01:48:44 -06:00
{ /* From NIST Special Publication 800-38B, AES-128 */
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = zeroed_string,
.digest = "\xbb\x1d\x69\x29\xe9\x59\x37\x28"
"\x7f\xa3\x7d\x12\x9b\x75\x67\x46",
.psize = 0,
.ksize = 16,
}, {
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a",
.digest = "\x07\x0a\x16\xb4\x6b\x4d\x41\x44"
"\xf7\x9b\xdd\x9d\xd0\x4a\x28\x7c",
.psize = 16,
.ksize = 16,
}, {
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11",
.digest = "\xdf\xa6\x67\x47\xde\x9a\xe6\x30"
"\x30\xca\x32\x61\x14\x97\xc8\x27",
.psize = 40,
.ksize = 16,
}, {
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.digest = "\x51\xf0\xbe\xbf\x7e\x3b\x9d\x92"
"\xfc\x49\x74\x17\x79\x36\x3c\xfe",
.psize = 64,
.ksize = 16,
}, { /* From NIST Special Publication 800-38B, AES-256 */
.key = "\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.plaintext = zeroed_string,
.digest = "\x02\x89\x62\xf6\x1b\x7b\xf8\x9e"
"\xfc\x6b\x55\x1f\x46\x67\xd9\x83",
.psize = 0,
.ksize = 32,
}, {
.key = "\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.digest = "\xe1\x99\x21\x90\x54\x9f\x6e\xd5"
"\x69\x6a\x2c\x05\x6c\x31\x54\x10",
.psize = 64,
.ksize = 32,
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec aes_cbcmac_tv_template[] = {
crypto: testmgr - add test cases for cbcmac(aes) In preparation of splitting off the CBC-MAC transform in the CCM driver into a separate algorithm, define some test cases for the AES incarnation of cbcmac. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-03 07:49:35 -07:00
{
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a",
.digest = "\x3a\xd7\x7b\xb4\x0d\x7a\x36\x60"
"\xa8\x9e\xca\xf3\x24\x66\xef\x97",
.psize = 16,
.ksize = 16,
}, {
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30",
.digest = "\x9d\x0d\xd0\x63\xfb\xcb\x24\x43"
"\xf8\xf2\x76\x03\xac\x39\xb0\x9d",
.psize = 33,
.ksize = 16,
.np = 2,
.tap = { 7, 26 },
}, {
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37",
.digest = "\xc0\x71\x73\xb8\xa0\x2c\x11\x7c"
"\xaf\xdc\xb2\xf8\x89\x32\xa3\x3a",
.psize = 63,
.ksize = 16,
}, {
.key = "\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10"
"\x1c",
.digest = "\x6a\x4e\xdb\x21\x47\x51\xdf\x4f"
"\xa8\x4d\x4c\x10\x3b\x72\x7d\xd6",
.psize = 65,
.ksize = 32,
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec des3_ede_cmac64_tv_template[] = {
crypto: add CMAC support to CryptoAPI Patch adds support for NIST recommended block cipher mode CMAC to CryptoAPI. This work is based on Tom St Denis' earlier patch, http://marc.info/?l=linux-crypto-vger&m=135877306305466&w=2 Cc: Tom St Denis <tstdenis@elliptictech.com> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-08 01:48:44 -06:00
/*
* From NIST Special Publication 800-38B, Three Key TDEA
* Corrected test vectors from:
* http://csrc.nist.gov/publications/nistpubs/800-38B/Updated_CMAC_Examples.pdf
*/
{
.key = "\x8a\xa8\x3b\xf8\xcb\xda\x10\x62"
"\x0b\xc1\xbf\x19\xfb\xb6\xcd\x58"
"\xbc\x31\x3d\x4a\x37\x1c\xa8\xb5",
.plaintext = zeroed_string,
.digest = "\xb7\xa6\x88\xe1\x22\xff\xaf\x95",
.psize = 0,
.ksize = 24,
}, {
.key = "\x8a\xa8\x3b\xf8\xcb\xda\x10\x62"
"\x0b\xc1\xbf\x19\xfb\xb6\xcd\x58"
"\xbc\x31\x3d\x4a\x37\x1c\xa8\xb5",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96",
.digest = "\x8e\x8f\x29\x31\x36\x28\x37\x97",
.psize = 8,
.ksize = 24,
}, {
.key = "\x8a\xa8\x3b\xf8\xcb\xda\x10\x62"
"\x0b\xc1\xbf\x19\xfb\xb6\xcd\x58"
"\xbc\x31\x3d\x4a\x37\x1c\xa8\xb5",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57",
.digest = "\x74\x3d\xdb\xe0\xce\x2d\xc2\xed",
.psize = 20,
.ksize = 24,
}, {
.key = "\x8a\xa8\x3b\xf8\xcb\xda\x10\x62"
"\x0b\xc1\xbf\x19\xfb\xb6\xcd\x58"
"\xbc\x31\x3d\x4a\x37\x1c\xa8\xb5",
.plaintext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51",
.digest = "\x33\xe6\xb1\x09\x24\x00\xea\xe5",
.psize = 32,
.ksize = 24,
}
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec aes_xcbc128_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = zeroed_string,
.digest = "\x75\xf0\x25\x1d\x52\x8a\xc0\x1c"
"\x45\x73\xdf\xd5\x84\xd7\x9f\x29",
.psize = 0,
.ksize = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = "\x00\x01\x02",
.digest = "\x5b\x37\x65\x80\xae\x2f\x19\xaf"
"\xe7\x21\x9c\xee\xf1\x72\x75\x6f",
.psize = 3,
.ksize = 16,
} , {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.digest = "\xd2\xa2\x46\xfa\x34\x9b\x68\xa7"
"\x99\x98\xa4\x39\x4f\xf7\xa2\x63",
.psize = 16,
.ksize = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13",
.digest = "\x47\xf5\x1b\x45\x64\x96\x62\x15"
"\xb8\x98\x5c\x63\x05\x5e\xd3\x08",
.tap = { 10, 10 },
.psize = 20,
.np = 2,
.ksize = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.digest = "\xf5\x4f\x0e\xc8\xd2\xb9\xf3\xd3"
"\x68\x07\x73\x4b\xd5\x28\x3f\xd4",
.psize = 32,
.ksize = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.plaintext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21",
.digest = "\xbe\xcb\xb3\xbc\xcd\xb5\x18\xa3"
"\x06\x77\xd5\x48\x1f\xb6\xb4\xd8",
.tap = { 17, 17 },
.psize = 34,
.np = 2,
.ksize = 16,
}
};
crypto: vmac - add nonced version with big endian digest Currently the VMAC template uses a "nonce" hardcoded to 0, which makes it insecure unless a unique key is set for every message. Also, the endianness of the final digest is wrong: the implementation uses little endian, but the VMAC specification has it as big endian, as do other VMAC implementations such as the one in Crypto++. Add a new VMAC template where the nonce is passed as the first 16 bytes of data (similar to what is done for Poly1305's nonce), and the digest is big endian. Call it "vmac64", since the old name of simply "vmac" didn't clarify whether the implementation is of VMAC-64 or of VMAC-128 (which produce 64-bit and 128-bit digests respectively); so we fix the naming ambiguity too. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-06-18 11:22:39 -06:00
static const char vmac64_string1[144] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\x01', '\x01', '\x01', '\x01', '\x02', '\x03', '\x02', '\x02',
'\x02', '\x04', '\x01', '\x07', '\x04', '\x01', '\x04', '\x03',
};
static const char vmac64_string2[144] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'a', 'b', 'c',
};
static const char vmac64_string3[144] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'a', 'b', 'c', 'a', 'b', 'c', 'a', 'b',
'c', 'a', 'b', 'c', 'a', 'b', 'c', 'a',
'b', 'c', 'a', 'b', 'c', 'a', 'b', 'c',
'a', 'b', 'c', 'a', 'b', 'c', 'a', 'b',
'c', 'a', 'b', 'c', 'a', 'b', 'c', 'a',
'b', 'c', 'a', 'b', 'c', 'a', 'b', 'c',
};
static const char vmac64_string4[33] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'b', 'c', 'e', 'f', 'i', 'j', 'l', 'm',
'o', 'p', 'r', 's', 't', 'u', 'w', 'x',
'z',
};
static const char vmac64_string5[143] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'r', 'm', 'b', 't', 'c', 'o', 'l', 'k',
']', '%', '9', '2', '7', '!', 'A',
};
static const char vmac64_string6[145] = {
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'\0', '\0', '\0', '\0', '\0', '\0', '\0', '\0',
'p', 't', '*', '7', 'l', 'i', '!', '#',
'w', '0', 'z', '/', '4', 'A', 'n',
};
static const struct hash_testvec vmac64_aes_tv_template[] = {
{ /* draft-krovetz-vmac-01 test vector 1 */
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = "\0\0\0\0\0\0\0\0bcdefghi",
.psize = 16,
.digest = "\x25\x76\xbe\x1c\x56\xd8\xb8\x1b",
}, { /* draft-krovetz-vmac-01 test vector 2 */
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = "\0\0\0\0\0\0\0\0bcdefghiabc",
.psize = 19,
.digest = "\x2d\x37\x6c\xf5\xb1\x81\x3c\xe5",
}, { /* draft-krovetz-vmac-01 test vector 3 */
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = "\0\0\0\0\0\0\0\0bcdefghi"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc",
.psize = 64,
.digest = "\xe8\x42\x1f\x61\xd5\x73\xd2\x98",
}, { /* draft-krovetz-vmac-01 test vector 4 */
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = "\0\0\0\0\0\0\0\0bcdefghi"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabcabc"
"abcabcabcabcabcabcabcabcabcabcabcabcabcabcabc",
.psize = 316,
.digest = "\x44\x92\xdf\x6c\x5c\xac\x1b\xbe",
.tap = { 1, 100, 200, 15 },
.np = 4,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ksize = 16,
.plaintext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.psize = 16,
.digest = "\x54\x7b\xa4\x77\x35\x80\x58\x07",
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ksize = 16,
.plaintext = vmac64_string1,
.psize = sizeof(vmac64_string1),
.digest = "\xa1\x8c\x68\xae\xd3\x3c\xf5\xce",
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ksize = 16,
.plaintext = vmac64_string2,
.psize = sizeof(vmac64_string2),
.digest = "\x2d\x14\xbd\x81\x73\xb0\x27\xc9",
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ksize = 16,
.plaintext = vmac64_string3,
.psize = sizeof(vmac64_string3),
.digest = "\x19\x0b\x47\x98\x8c\x95\x1a\x8d",
}, {
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.psize = 16,
.digest = "\x84\x8f\x55\x9e\x26\xa1\x89\x3b",
}, {
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = vmac64_string1,
.psize = sizeof(vmac64_string1),
.digest = "\xc2\x74\x8d\xf6\xb0\xab\x5e\xab",
}, {
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = vmac64_string2,
.psize = sizeof(vmac64_string2),
.digest = "\xdf\x09\x7b\x3d\x42\x68\x15\x11",
}, {
.key = "abcdefghijklmnop",
.ksize = 16,
.plaintext = vmac64_string3,
.psize = sizeof(vmac64_string3),
.digest = "\xd4\xfa\x8f\xed\xe1\x8f\x32\x8b",
}, {
.key = "a09b5cd!f#07K\x00\x00\x00",
.ksize = 16,
.plaintext = vmac64_string4,
.psize = sizeof(vmac64_string4),
.digest = "\x5f\xa1\x4e\x42\xea\x0f\xa5\xab",
}, {
.key = "a09b5cd!f#07K\x00\x00\x00",
.ksize = 16,
.plaintext = vmac64_string5,
.psize = sizeof(vmac64_string5),
.digest = "\x60\x67\xe8\x1d\xbc\x98\x31\x25",
}, {
.key = "a09b5cd!f#07K\x00\x00\x00",
.ksize = 16,
.plaintext = vmac64_string6,
.psize = sizeof(vmac64_string6),
.digest = "\x41\xeb\x65\x95\x47\x9b\xae\xc4",
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* SHA384 HMAC test vectors from RFC4231
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha384_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\xaf\xd0\x39\x44\xd8\x48\x95\x62"
"\x6b\x08\x25\xf4\xab\x46\x90\x7f"
"\x15\xf9\xda\xdb\xe4\x10\x1e\xc6"
"\x82\xaa\x03\x4c\x7c\xeb\xc5\x9c"
"\xfa\xea\x9e\xa9\x07\x6e\xde\x7f"
"\x4a\xf1\x52\xe8\xb2\xfa\x9c\xb6",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\xaf\x45\xd2\xe3\x76\x48\x40\x31"
"\x61\x7f\x78\xd2\xb5\x8a\x6b\x1b"
"\x9c\x7e\xf4\x64\xf5\xa0\x1b\x47"
"\xe4\x2e\xc3\x73\x63\x22\x44\x5e"
"\x8e\x22\x40\xca\x5e\x69\xe2\xc7"
"\x8b\x32\x39\xec\xfa\xb2\x16\x49",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Larger Than Block-Siz"
"e Key - Hash Key First",
.psize = 54,
.digest = "\x4e\xce\x08\x44\x85\x81\x3e\x90"
"\x88\xd2\xc6\x3a\x04\x1b\xc5\xb4"
"\x4f\x9e\xf1\x01\x2a\x2b\x58\x8f"
"\x3c\xd1\x1f\x05\x03\x3a\xc4\xc6"
"\x0c\x2e\xf6\xab\x40\x30\xfe\x82"
"\x96\x24\x8d\xf1\x63\xf4\x49\x52",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\x66\x17\x17\x8e\x94\x1f\x02\x0d"
"\x35\x1e\x2f\x25\x4e\x8f\xd3\x2c"
"\x60\x24\x20\xfe\xb0\xb8\xfb\x9a"
"\xdc\xce\xbb\x82\x46\x1e\x99\xc5"
"\xa6\x78\xcc\x31\xe7\x99\x17\x6d"
"\x38\x60\xe6\x11\x0c\x46\x52\x3e",
},
};
/*
* SHA512 HMAC test vectors from RFC4231
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha512_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x87\xaa\x7c\xde\xa5\xef\x61\x9d"
"\x4f\xf0\xb4\x24\x1a\x1d\x6c\xb0"
"\x23\x79\xf4\xe2\xce\x4e\xc2\x78"
"\x7a\xd0\xb3\x05\x45\xe1\x7c\xde"
"\xda\xa8\x33\xb7\xd6\xb8\xa7\x02"
"\x03\x8b\x27\x4e\xae\xa3\xf4\xe4"
"\xbe\x9d\x91\x4e\xeb\x61\xf1\x70"
"\x2e\x69\x6c\x20\x3a\x12\x68\x54",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x16\x4b\x7a\x7b\xfc\xf8\x19\xe2"
"\xe3\x95\xfb\xe7\x3b\x56\xe0\xa3"
"\x87\xbd\x64\x22\x2e\x83\x1f\xd6"
"\x10\x27\x0c\xd7\xea\x25\x05\x54"
"\x97\x58\xbf\x75\xc0\x5a\x99\x4a"
"\x6d\x03\x4f\x65\xf8\xf0\xe6\xfd"
"\xca\xea\xb1\xa3\x4d\x4a\x6b\x4b"
"\x63\x6e\x07\x0a\x38\xbc\xe7\x37",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Large"
"r Than Block-Siz"
"e Key - Hash Key"
" First",
.psize = 54,
.digest = "\x80\xb2\x42\x63\xc7\xc1\xa3\xeb"
"\xb7\x14\x93\xc1\xdd\x7b\xe8\xb4"
"\x9b\x46\xd1\xf4\x1b\x4a\xee\xc1"
"\x12\x1b\x01\x37\x83\xf8\xf3\x52"
"\x6b\x56\xd0\x37\xe0\x5f\x25\x98"
"\xbd\x0f\xd2\x21\x5d\x6a\x1e\x52"
"\x95\xe6\x4f\x73\xf6\x3f\x0a\xec"
"\x8b\x91\x5a\x98\x5d\x78\x65\x98",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext =
"This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\xe3\x7b\x6a\x77\x5d\xc8\x7d\xba"
"\xa4\xdf\xa9\xf9\x6e\x5e\x3f\xfd"
"\xde\xbd\x71\xf8\x86\x72\x89\x86"
"\x5d\xf5\xa3\x2d\x20\xcd\xc9\x44"
"\xb6\x02\x2c\xac\x3c\x49\x82\xb1"
"\x0d\x5e\xeb\x55\xc3\xe4\xde\x15"
"\x13\x46\x76\xfb\x6d\xe0\x44\x60"
"\x65\xc9\x74\x40\xfa\x8c\x6a\x58",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha3_224_tv_template[] = {
crypto: sha3 - Add HMAC-SHA3 test modes and test vectors This patch adds HMAC-SHA3 test modes in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-30 23:46:54 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x3b\x16\x54\x6b\xbc\x7b\xe2\x70"
"\x6a\x03\x1d\xca\xfd\x56\x37\x3d"
"\x98\x84\x36\x76\x41\xd8\xc5\x9a"
"\xf3\xc8\x60\xf7",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x7f\xdb\x8d\xd8\x8b\xd2\xf6\x0d"
"\x1b\x79\x86\x34\xad\x38\x68\x11"
"\xc2\xcf\xc8\x5b\xfa\xf5\xd5\x2b"
"\xba\xce\x5e\x66",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Large"
"r Than Block-Siz"
"e Key - Hash Key"
" First",
.psize = 54,
.digest = "\xb4\xa1\xf0\x4c\x00\x28\x7a\x9b"
"\x7f\x60\x75\xb3\x13\xd2\x79\xb8"
"\x33\xbc\x8f\x75\x12\x43\x52\xd0"
"\x5f\xb9\x99\x5f",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext =
"This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\x05\xd8\xcd\x6d\x00\xfa\xea\x8d"
"\x1e\xb6\x8a\xde\x28\x73\x0b\xbd"
"\x3c\xba\xb6\x92\x9f\x0a\x08\x6b"
"\x29\xcd\x62\xa0",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha3_256_tv_template[] = {
crypto: sha3 - Add HMAC-SHA3 test modes and test vectors This patch adds HMAC-SHA3 test modes in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-30 23:46:54 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\xba\x85\x19\x23\x10\xdf\xfa\x96"
"\xe2\xa3\xa4\x0e\x69\x77\x43\x51"
"\x14\x0b\xb7\x18\x5e\x12\x02\xcd"
"\xcc\x91\x75\x89\xf9\x5e\x16\xbb",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\xc7\xd4\x07\x2e\x78\x88\x77\xae"
"\x35\x96\xbb\xb0\xda\x73\xb8\x87"
"\xc9\x17\x1f\x93\x09\x5b\x29\x4a"
"\xe8\x57\xfb\xe2\x64\x5e\x1b\xa5",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Large"
"r Than Block-Siz"
"e Key - Hash Key"
" First",
.psize = 54,
.digest = "\xed\x73\xa3\x74\xb9\x6c\x00\x52"
"\x35\xf9\x48\x03\x2f\x09\x67\x4a"
"\x58\xc0\xce\x55\x5c\xfc\x1f\x22"
"\x3b\x02\x35\x65\x60\x31\x2c\x3b",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext =
"This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\x65\xc5\xb0\x6d\x4c\x3d\xe3\x2a"
"\x7a\xef\x87\x63\x26\x1e\x49\xad"
"\xb6\xe2\x29\x3e\xc8\xe7\xc6\x1e"
"\x8d\xe6\x17\x01\xfc\x63\xe1\x23",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha3_384_tv_template[] = {
crypto: sha3 - Add HMAC-SHA3 test modes and test vectors This patch adds HMAC-SHA3 test modes in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-30 23:46:54 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\x68\xd2\xdc\xf7\xfd\x4d\xdd\x0a"
"\x22\x40\xc8\xa4\x37\x30\x5f\x61"
"\xfb\x73\x34\xcf\xb5\xd0\x22\x6e"
"\x1b\xc2\x7d\xc1\x0a\x2e\x72\x3a"
"\x20\xd3\x70\xb4\x77\x43\x13\x0e"
"\x26\xac\x7e\x3d\x53\x28\x86\xbd",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\xf1\x10\x1f\x8c\xbf\x97\x66\xfd"
"\x67\x64\xd2\xed\x61\x90\x3f\x21"
"\xca\x9b\x18\xf5\x7c\xf3\xe1\xa2"
"\x3c\xa1\x35\x08\xa9\x32\x43\xce"
"\x48\xc0\x45\xdc\x00\x7f\x26\xa2"
"\x1b\x3f\x5e\x0e\x9d\xf4\xc2\x0a",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Large"
"r Than Block-Siz"
"e Key - Hash Key"
" First",
.psize = 54,
.digest = "\x0f\xc1\x95\x13\xbf\x6b\xd8\x78"
"\x03\x70\x16\x70\x6a\x0e\x57\xbc"
"\x52\x81\x39\x83\x6b\x9a\x42\xc3"
"\xd4\x19\xe4\x98\xe0\xe1\xfb\x96"
"\x16\xfd\x66\x91\x38\xd3\x3a\x11"
"\x05\xe0\x7c\x72\xb6\x95\x3b\xcc",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext =
"This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\x02\x6f\xdf\x6b\x50\x74\x1e\x37"
"\x38\x99\xc9\xf7\xd5\x40\x6d\x4e"
"\xb0\x9f\xc6\x66\x56\x36\xfc\x1a"
"\x53\x00\x29\xdd\xf5\xcf\x3c\xa5"
"\xa9\x00\xed\xce\x01\xf5\xf6\x1e"
"\x2f\x40\x8c\xdf\x2f\xd3\xe7\xe8",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec hmac_sha3_512_tv_template[] = {
crypto: sha3 - Add HMAC-SHA3 test modes and test vectors This patch adds HMAC-SHA3 test modes in tcrypt module and related test vectors. Signed-off-by: Raveendra Padasalagi <raveendra.padasalagi@broadcom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2016-06-30 23:46:54 -06:00
{
.key = "\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b\x0b\x0b\x0b\x0b"
"\x0b\x0b\x0b\x0b",
.ksize = 20,
.plaintext = "Hi There",
.psize = 8,
.digest = "\xeb\x3f\xbd\x4b\x2e\xaa\xb8\xf5"
"\xc5\x04\xbd\x3a\x41\x46\x5a\xac"
"\xec\x15\x77\x0a\x7c\xab\xac\x53"
"\x1e\x48\x2f\x86\x0b\x5e\xc7\xba"
"\x47\xcc\xb2\xc6\xf2\xaf\xce\x8f"
"\x88\xd2\x2b\x6d\xc6\x13\x80\xf2"
"\x3a\x66\x8f\xd3\x88\x8b\xb8\x05"
"\x37\xc0\xa0\xb8\x64\x07\x68\x9e",
}, {
.key = "Jefe",
.ksize = 4,
.plaintext = "what do ya want for nothing?",
.psize = 28,
.digest = "\x5a\x4b\xfe\xab\x61\x66\x42\x7c"
"\x7a\x36\x47\xb7\x47\x29\x2b\x83"
"\x84\x53\x7c\xdb\x89\xaf\xb3\xbf"
"\x56\x65\xe4\xc5\xe7\x09\x35\x0b"
"\x28\x7b\xae\xc9\x21\xfd\x7c\xa0"
"\xee\x7a\x0c\x31\xd0\x22\xa9\x5e"
"\x1f\xc9\x2b\xa9\xd7\x7d\xf8\x83"
"\x96\x02\x75\xbe\xb4\xe6\x20\x24",
.np = 4,
.tap = { 7, 7, 7, 7 }
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext = "Test Using Large"
"r Than Block-Siz"
"e Key - Hash Key"
" First",
.psize = 54,
.digest = "\x00\xf7\x51\xa9\xe5\x06\x95\xb0"
"\x90\xed\x69\x11\xa4\xb6\x55\x24"
"\x95\x1c\xdc\x15\xa7\x3a\x5d\x58"
"\xbb\x55\x21\x5e\xa2\xcd\x83\x9a"
"\xc7\x9d\x2b\x44\xa3\x9b\xaf\xab"
"\x27\xe8\x3f\xde\x9e\x11\xf6\x34"
"\x0b\x11\xd9\x91\xb1\xb9\x1b\xf2"
"\xee\xe7\xfc\x87\x24\x26\xc3\xa4",
}, {
.key = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa",
.ksize = 131,
.plaintext =
"This is a test u"
"sing a larger th"
"an block-size ke"
"y and a larger t"
"han block-size d"
"ata. The key nee"
"ds to be hashed "
"before being use"
"d by the HMAC al"
"gorithm.",
.psize = 152,
.digest = "\x38\xa4\x56\xa0\x04\xbd\x10\xd3"
"\x2c\x9a\xb8\x33\x66\x84\x11\x28"
"\x62\xc3\xdb\x61\xad\xcc\xa3\x18"
"\x29\x35\x5e\xaf\x46\xfd\x5c\x73"
"\xd0\x6a\x1f\x0d\x13\xfe\xc9\xa6"
"\x52\xfb\x38\x11\xb5\x77\xb1\xb1"
"\xd1\xb9\x78\x9f\x97\xae\x5b\x83"
"\xc6\xf4\x4d\xfc\xf1\xd6\x7e\xba",
},
};
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
/*
* Poly1305 test vectors from RFC7539 A.3.
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec poly1305_tv_template[] = {
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
{ /* Test Vector #1 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 96,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #2 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x36\xe5\xf6\xb5\xc5\xe0\x60\x70"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\xf0\xef\xca\x96\x22\x7a\x86\x3e"
"\x41\x6e\x79\x20\x73\x75\x62\x6d"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x69\x73\x73\x69\x6f\x6e\x20\x74"
"\x6f\x20\x74\x68\x65\x20\x49\x45"
"\x54\x46\x20\x69\x6e\x74\x65\x6e"
"\x64\x65\x64\x20\x62\x79\x20\x74"
"\x68\x65\x20\x43\x6f\x6e\x74\x72"
"\x69\x62\x75\x74\x6f\x72\x20\x66"
"\x6f\x72\x20\x70\x75\x62\x6c\x69"
"\x63\x61\x74\x69\x6f\x6e\x20\x61"
"\x73\x20\x61\x6c\x6c\x20\x6f\x72"
"\x20\x70\x61\x72\x74\x20\x6f\x66"
"\x20\x61\x6e\x20\x49\x45\x54\x46"
"\x20\x49\x6e\x74\x65\x72\x6e\x65"
"\x74\x2d\x44\x72\x61\x66\x74\x20"
"\x6f\x72\x20\x52\x46\x43\x20\x61"
"\x6e\x64\x20\x61\x6e\x79\x20\x73"
"\x74\x61\x74\x65\x6d\x65\x6e\x74"
"\x20\x6d\x61\x64\x65\x20\x77\x69"
"\x74\x68\x69\x6e\x20\x74\x68\x65"
"\x20\x63\x6f\x6e\x74\x65\x78\x74"
"\x20\x6f\x66\x20\x61\x6e\x20\x49"
"\x45\x54\x46\x20\x61\x63\x74\x69"
"\x76\x69\x74\x79\x20\x69\x73\x20"
"\x63\x6f\x6e\x73\x69\x64\x65\x72"
"\x65\x64\x20\x61\x6e\x20\x22\x49"
"\x45\x54\x46\x20\x43\x6f\x6e\x74"
"\x72\x69\x62\x75\x74\x69\x6f\x6e"
"\x22\x2e\x20\x53\x75\x63\x68\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x63\x6c\x75"
"\x64\x65\x20\x6f\x72\x61\x6c\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x20\x49\x45"
"\x54\x46\x20\x73\x65\x73\x73\x69"
"\x6f\x6e\x73\x2c\x20\x61\x73\x20"
"\x77\x65\x6c\x6c\x20\x61\x73\x20"
"\x77\x72\x69\x74\x74\x65\x6e\x20"
"\x61\x6e\x64\x20\x65\x6c\x65\x63"
"\x74\x72\x6f\x6e\x69\x63\x20\x63"
"\x6f\x6d\x6d\x75\x6e\x69\x63\x61"
"\x74\x69\x6f\x6e\x73\x20\x6d\x61"
"\x64\x65\x20\x61\x74\x20\x61\x6e"
"\x79\x20\x74\x69\x6d\x65\x20\x6f"
"\x72\x20\x70\x6c\x61\x63\x65\x2c"
"\x20\x77\x68\x69\x63\x68\x20\x61"
"\x72\x65\x20\x61\x64\x64\x72\x65"
"\x73\x73\x65\x64\x20\x74\x6f",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 407,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x36\xe5\xf6\xb5\xc5\xe0\x60\x70"
"\xf0\xef\xca\x96\x22\x7a\x86\x3e",
}, { /* Test Vector #3 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x36\xe5\xf6\xb5\xc5\xe0\x60\x70"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xf0\xef\xca\x96\x22\x7a\x86\x3e"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x41\x6e\x79\x20\x73\x75\x62\x6d"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x69\x73\x73\x69\x6f\x6e\x20\x74"
"\x6f\x20\x74\x68\x65\x20\x49\x45"
"\x54\x46\x20\x69\x6e\x74\x65\x6e"
"\x64\x65\x64\x20\x62\x79\x20\x74"
"\x68\x65\x20\x43\x6f\x6e\x74\x72"
"\x69\x62\x75\x74\x6f\x72\x20\x66"
"\x6f\x72\x20\x70\x75\x62\x6c\x69"
"\x63\x61\x74\x69\x6f\x6e\x20\x61"
"\x73\x20\x61\x6c\x6c\x20\x6f\x72"
"\x20\x70\x61\x72\x74\x20\x6f\x66"
"\x20\x61\x6e\x20\x49\x45\x54\x46"
"\x20\x49\x6e\x74\x65\x72\x6e\x65"
"\x74\x2d\x44\x72\x61\x66\x74\x20"
"\x6f\x72\x20\x52\x46\x43\x20\x61"
"\x6e\x64\x20\x61\x6e\x79\x20\x73"
"\x74\x61\x74\x65\x6d\x65\x6e\x74"
"\x20\x6d\x61\x64\x65\x20\x77\x69"
"\x74\x68\x69\x6e\x20\x74\x68\x65"
"\x20\x63\x6f\x6e\x74\x65\x78\x74"
"\x20\x6f\x66\x20\x61\x6e\x20\x49"
"\x45\x54\x46\x20\x61\x63\x74\x69"
"\x76\x69\x74\x79\x20\x69\x73\x20"
"\x63\x6f\x6e\x73\x69\x64\x65\x72"
"\x65\x64\x20\x61\x6e\x20\x22\x49"
"\x45\x54\x46\x20\x43\x6f\x6e\x74"
"\x72\x69\x62\x75\x74\x69\x6f\x6e"
"\x22\x2e\x20\x53\x75\x63\x68\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x63\x6c\x75"
"\x64\x65\x20\x6f\x72\x61\x6c\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x20\x49\x45"
"\x54\x46\x20\x73\x65\x73\x73\x69"
"\x6f\x6e\x73\x2c\x20\x61\x73\x20"
"\x77\x65\x6c\x6c\x20\x61\x73\x20"
"\x77\x72\x69\x74\x74\x65\x6e\x20"
"\x61\x6e\x64\x20\x65\x6c\x65\x63"
"\x74\x72\x6f\x6e\x69\x63\x20\x63"
"\x6f\x6d\x6d\x75\x6e\x69\x63\x61"
"\x74\x69\x6f\x6e\x73\x20\x6d\x61"
"\x64\x65\x20\x61\x74\x20\x61\x6e"
"\x79\x20\x74\x69\x6d\x65\x20\x6f"
"\x72\x20\x70\x6c\x61\x63\x65\x2c"
"\x20\x77\x68\x69\x63\x68\x20\x61"
"\x72\x65\x20\x61\x64\x64\x72\x65"
"\x73\x73\x65\x64\x20\x74\x6f",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 407,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\xf3\x47\x7e\x7c\xd9\x54\x17\xaf"
"\x89\xa6\xb8\x79\x4c\x31\x0c\xf0",
}, { /* Test Vector #4 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0"
"\x27\x54\x77\x61\x73\x20\x62\x72"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x69\x6c\x6c\x69\x67\x2c\x20\x61"
"\x6e\x64\x20\x74\x68\x65\x20\x73"
"\x6c\x69\x74\x68\x79\x20\x74\x6f"
"\x76\x65\x73\x0a\x44\x69\x64\x20"
"\x67\x79\x72\x65\x20\x61\x6e\x64"
"\x20\x67\x69\x6d\x62\x6c\x65\x20"
"\x69\x6e\x20\x74\x68\x65\x20\x77"
"\x61\x62\x65\x3a\x0a\x41\x6c\x6c"
"\x20\x6d\x69\x6d\x73\x79\x20\x77"
"\x65\x72\x65\x20\x74\x68\x65\x20"
"\x62\x6f\x72\x6f\x67\x6f\x76\x65"
"\x73\x2c\x0a\x41\x6e\x64\x20\x74"
"\x68\x65\x20\x6d\x6f\x6d\x65\x20"
"\x72\x61\x74\x68\x73\x20\x6f\x75"
"\x74\x67\x72\x61\x62\x65\x2e",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 159,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x45\x41\x66\x9a\x7e\xaa\xee\x61"
"\xe7\x08\xdc\x7c\xbc\xc5\xeb\x62",
}, { /* Test Vector #5 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x02\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xff\xff\xff\xff\xff\xff\xff\xff"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 48,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x03\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #6 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x02\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xff\xff\xff\xff\xff\xff\xff\xff"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\x02\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 48,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x03\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #7 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x01\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xff\xff\xff\xff\xff\xff\xff\xff"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xf0\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\x11\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 80,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x05\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #8 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x01\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xfb\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 80,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #9 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x02\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xfd\xff\xff\xff\xff\xff\xff\xff"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\xff\xff\xff\xff\xff\xff\xff\xff",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 48,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\xfa\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff",
}, { /* Test Vector #10 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x01\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x04\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xe3\x35\x94\xd7\x50\x5e\x43\xb9"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x33\x94\xd7\x50\x5e\x43\x79\xcd"
"\x01\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x01\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 96,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x14\x00\x00\x00\x00\x00\x00\x00"
"\x55\x00\x00\x00\x00\x00\x00\x00",
}, { /* Test Vector #11 */
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.plaintext = "\x01\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x04\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\xe3\x35\x94\xd7\x50\x5e\x43\xb9"
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x33\x94\xd7\x50\x5e\x43\x79\xcd"
"\x01\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: poly1305 - Pass key as first two message blocks to each desc_ctx The Poly1305 authenticator requires a unique key for each generated tag. This implies that we can't set the key per tfm, as multiple users set individual keys. Instead we pass a desc specific key as the first two blocks of the message to authenticate in update(). Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-16 03:34:16 -06:00
.psize = 80,
crypto: testmgr - Add Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:59 -06:00
.digest = "\x13\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* DES test vectors.
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* From Applied Cryptography */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7",
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Same key, different plaintext block */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x22\x33\x44\x55\x66\x77\x88\x99",
.ctext = "\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Sbox test from NBS */
.key = "\x7c\xa1\x10\x45\x4a\x1a\x6e\x57",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\xa1\xd6\xd0\x39\x77\x67\x42",
.ctext = "\x69\x0f\x5b\x0d\x9a\x26\x93\x9b",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Three blocks */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xca\xfe\xba\xbe\xfe\xed\xbe\xef",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b"
"\xb4\x99\x26\xf7\x1f\xe1\xd4\x90",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 24,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Weak key */
crypto: testmgr - Wrap the LHS in expressions of the form !x == y In the test manager, there are a number of if-statements with expressions of the form !x == y that incur warnings with gcc-5 of the following form: ../crypto/testmgr.c: In function '__test_aead': ../crypto/testmgr.c:523:12: warning: logical not is only applied to the left hand side of comparison [-Wlogical-not-parentheses] if (!ret == template[i].fail) { ^ By converting the 'fail' member of struct aead_testvec and struct cipher_testvec to a bool, we can get rid of the warnings. Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-04-28 08:36:36 -06:00
.fail = true,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.wk = 1,
.key = "\x01\x01\x01\x01\x01\x01\x01\x01",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7",
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Two blocks -- for testing encryption across pages */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x33\x44\x55\x66\x77\x88\x99",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 2,
.tap = { 8, 8 }
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
"\xa3\x99\x7b\xca\xaf\x69\xa0\xf5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
"\x69\x0f\x5b\x0d\x9a\x26\x93\x9b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
.np = 2,
.tap = { 8, 8 }
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
"\xa3\x99\x7b\xca\xaf\x69\xa0\xf5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
"\x69\x0f\x5b\x0d\x9a\x26\x93\x9b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: testmgr - add extra ecb(des) encryption test vectors Two "ecb(des)" decryption test vectors don't exactly match any of the encryption test vectors with input and result swapped. In preparation for removing the decryption test vectors, add these to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:25 -06:00
.np = 3,
.tap = { 3, 12, 1 }
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Four blocks -- for testing encryption with chunking */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xca\xfe\xba\xbe\xfe\xed\xbe\xef"
"\x22\x33\x44\x55\x66\x77\x88\x99",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b"
"\xb4\x99\x26\xf7\x1f\xe1\xd4\x90"
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 3,
.tap = { 14, 10, 8 }
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xca\xfe\xba\xbe\xfe\xed\xbe\xef",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b"
"\xb4\x99\x26\xf7\x1f\xe1\xd4\x90",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 24,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 4,
.tap = { 2, 1, 3, 18 }
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x33\x44\x55\x66\x77\x88\x99",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xf7\x9c\x89\x2a\x33\x8f\x4a\x8b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 5,
.tap = { 2, 2, 2, 2, 8 }
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xe7",
.ctext = "\xc9\x57\x44\x25\x6a\x5e\xd3\x1d",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 8,
.tap = { 1, 1, 1, 1, 1, 1, 1, 1 }
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
}, { /* Generated with Crypto++ */
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55",
.klen = 8,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x88\xCB\x1F\xAB\x2F\x2A\x49\x57"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\x92\xB9\x77\xFF\x2F\x47\x58\xDD"
"\xD7\x8A\x91\x95\x26\x33\x78\xB2"
"\x33\xBA\xB2\x3E\x02\xF5\x1F\xEF"
"\x98\xC5\xA6\xD2\x7D\x79\xEC\xB3"
"\x45\xF3\x4C\x61\xAC\x6C\xC2\x55"
"\xE5\xD3\x06\x58\x8A\x42\x3E\xDD"
"\x3D\x20\x45\xE9\x6F\x0D\x25\xA8"
"\xA5\xC7\x69\xCE\xD5\x3B\x7B\xC9"
"\x9E\x65\xE7\xA3\xF2\xE4\x18\x94"
"\xD2\x81\xE9\x33\x2B\x2D\x49\xC4"
"\xFE\xDA\x7F\xE2\xF2\x8C\x9C\xDC"
"\x73\x58\x11\x1F\x81\xD7\x21\x1A"
"\x80\xD0\x0D\xE8\x45\xD6\xD8\xD5"
"\x2E\x51\x16\xCA\x09\x89\x54\x62"
"\xF7\x04\x3D\x75\xB9\xA3\x84\xF4"
"\x62\xF0\x02\x58\x83\xAF\x30\x87"
"\x85\x3F\x01\xCD\x8E\x58\x42\xC4"
"\x41\x73\xE0\x15\x0A\xE6\x2E\x80"
"\x94\xF8\x5B\x3A\x4E\xDF\x51\xB2"
"\x9D\xE4\xC4\x9D\xF7\x3F\xF8\x8E"
"\x37\x22\x4D\x00\x2A\xEF\xC1\x0F"
"\x14\xA0\x66\xAB\x79\x39\xD0\x8E"
"\xE9\x95\x61\x74\x12\xED\x07\xD7"
"\xDD\x95\xDC\x7B\x57\x25\x27\x9C"
"\x51\x96\x16\xF7\x94\x61\xB8\x87"
"\xF0\x21\x1B\x32\xFB\x07\x0F\x29"
"\x56\xBD\x9D\x22\xA2\x9F\xA2\xB9"
"\x46\x31\x4C\x5E\x2E\x95\x61\xEF"
"\xE1\x58\x39\x09\xB4\x8B\x40\xAC"
"\x5F\x62\xC7\x72\xD9\xFC\xCB\x9A",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 248,
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 248 - 10, 2, 8 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des_cbc_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* From OpenSSL */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.iv = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x37\x36\x35\x34\x33\x32\x31\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x4e\x6f\x77\x20\x69\x73\x20\x74"
"\x68\x65\x20\x74\x69\x6d\x65\x20",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xcc\xd1\x73\xff\xab\x20\x39\xf4"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xac\xd8\xae\xfd\xdf\xd8\xa1\xeb"
"\x46\x8e\x91\x15\x78\x88\xba\x68",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 24,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* FIPS Pub 81 */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.iv = "\x12\x34\x56\x78\x90\xab\xcd\xef",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x4e\x6f\x77\x20\x69\x73\x20\x74",
.ctext = "\xe5\xc7\xcd\xde\x87\x2b\xf2\x7c",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.iv = "\xe5\xc7\xcd\xde\x87\x2b\xf2\x7c",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x68\x65\x20\x74\x69\x6d\x65\x20",
.ctext = "\x43\xe9\x34\x00\x8c\x38\x9c\x0f",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.iv = "\x43\xe9\x34\x00\x8c\x38\x9c\x0f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x66\x6f\x72\x20\x61\x6c\x6c\x20",
.ctext = "\x68\x37\x88\x49\x9a\x7c\x05\xf6",
.len = 8,
crypto: testmgr - make an cbc(des) encryption test vector chunked One "cbc(des)" decryption test vector doesn't exactly match an encryption test vector with input and result swapped. It's *almost* the same as one, but the decryption version is "chunked" while the encryption version is "unchunked". In preparation for removing the decryption test vectors, make the encryption one both chunked and unchunked, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:26 -06:00
.np = 2,
.tap = { 4, 4 },
.also_non_np = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Copy of openssl vector for chunk testing */
/* From OpenSSL */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.iv = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x37\x36\x35\x34\x33\x32\x31\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x4e\x6f\x77\x20\x69\x73\x20\x74"
"\x68\x65\x20\x74\x69\x6d\x65\x20",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xcc\xd1\x73\xff\xab\x20\x39\xf4"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xac\xd8\xae\xfd\xdf\xd8\xa1\xeb"
"\x46\x8e\x91\x15\x78\x88\xba\x68",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 24,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 2,
.tap = { 13, 11 }
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
}, { /* Generated with Crypto++ */
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55",
.klen = 8,
.iv = "\xE7\x82\x1D\xB8\x53\x11\xAC\x47",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x71\xCC\x56\x1C\x87\x2C\x43\x20"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\x1C\x20\x13\x09\xF9\x2B\x40\x47"
"\x99\x10\xD1\x1B\x65\x33\x33\xBA"
"\x88\x0D\xA2\xD1\x86\xFF\x4D\xF4"
"\x5A\x0C\x12\x96\x32\x57\xAA\x26"
"\xA7\xF4\x32\x8D\xBC\x10\x31\x9E"
"\x81\x72\x74\xDE\x30\x19\x69\x49"
"\x54\x9C\xC3\xEB\x0B\x97\xDD\xD1"
"\xE8\x6D\x0D\x05\x83\xA5\x12\x08"
"\x47\xF8\x88\x03\x86\x51\x3C\xEF"
"\xE7\x11\x73\x4D\x44\x2B\xE2\x16"
"\xE8\xA5\x06\x50\x66\x70\x0E\x14"
"\xBA\x21\x3B\xD5\x23\x5B\xA7\x8F"
"\x56\xB6\xA7\x44\xDB\x86\xAB\x69"
"\x33\x3C\xBE\x64\xC4\x22\xD3\xFE"
"\x49\x90\x88\x6A\x09\x8F\x76\x59"
"\xCB\xB7\xA0\x2D\x79\x75\x92\x8A"
"\x82\x1D\xC2\xFE\x09\x1F\x78\x6B"
"\x2F\xD6\xA4\x87\x1E\xC4\x53\x63"
"\x80\x02\x61\x2F\xE3\x46\xB6\xB5"
"\xAA\x95\xF4\xEE\xA7\x64\x2B\x4F"
"\x20\xCF\xD2\x47\x4E\x39\x65\xB3"
"\x11\x87\xA2\x6C\x49\x7E\x36\xC7"
"\x62\x8B\x48\x0D\x6A\x64\x00\xBD"
"\x71\x91\x8C\xE9\x70\x19\x01\x4F"
"\x4E\x68\x23\xBA\xDA\x24\x2E\x45"
"\x02\x14\x33\x21\xAE\x58\x4B\xCF"
"\x3B\x4B\xE8\xF8\xF6\x4F\x34\x93"
"\xD7\x07\x8A\xD7\x18\x92\x36\x8C"
"\x82\xA9\xBD\x6A\x31\x91\x39\x11"
"\xC6\x4A\xF3\x55\xC7\x29\x2E\x63",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 248,
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 248 - 10, 2, 8 },
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des_ctr_tv_template[] = {
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
{ /* Generated with Crypto++ */
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55",
.klen = 8,
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x2F\x96\x06\x0F\x50\xC9\x68\x03"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\x0F\x31\xD4\x64\xA5\x29\x77\x35"
"\xBC\x7A\x9F\x19\xE7\x0D\x33\x3E"
"\x12\x0B\x8C\xAE\x48\xAE\xD9\x02"
"\x0A\xD4\xB0\xD6\x37\xB2\x65\x1C"
"\x4B\x65\xEB\x24\xB5\x8E\xAD\x47"
"\x0D\xDA\x79\x77\xA0\x29\xA0\x2B"
"\xC8\x0F\x85\xDC\x03\x13\xA9\x04"
"\x19\x40\xBE\xBE\x5C\x49\x4A\x69"
"\xED\xE8\xE1\x9E\x14\x43\x74\xDE"
"\xEC\x6E\x11\x3F\x36\xEF\x7B\xFB"
"\xBE\x4C\x91\x43\x22\x65\x72\x48"
"\xE2\x12\xED\x88\xAC\xA7\xC9\x91"
"\x14\xA2\x36\x1C\x29\xFF\xC8\x4F"
"\x72\x5C\x4B\xB0\x1E\x93\xC2\xFA"
"\x9D\x53\x86\xA0\xAE\xC6\xB7\x3C"
"\x59\x0C\xD0\x8F\xA6\xD8\xA4\x31"
"\xB7\x30\x1C\x21\x38\xFB\x68\x8C"
"\x2E\xF5\x6E\x73\xC3\x16\x5F\x12"
"\x0C\x33\xB9\x1E\x7B\x70\xDE\x86"
"\x32\xB3\xC1\x16\xAB\xD9\x49\x0B"
"\x96\x28\x72\x6B\xF3\x30\xA9\xEB"
"\x69\xE2\x1E\x58\x46\xA2\x8E\xC7"
"\xC0\xEF\x07\xB7\x77\x2C\x00\x05"
"\x46\xBD\xFE\x53\x81\x8B\xA4\x03"
"\x20\x0F\xDB\x78\x0B\x1F\x53\x04"
"\x4C\x60\x4C\xC3\x2A\x86\x86\x7E"
"\x13\xD2\x26\xED\x5D\x3E\x9C\xF2"
"\x5C\xC4\x15\xC9\x9A\x21\xC5\xCD"
"\x19\x7F\x99\x19\x53\xCE\x1D\x14"
"\x69\x74\xA1\x06\x46\x0F\x4E\x75",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 248,
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 248 - 10, 2, 8 },
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
}, { /* Generated with Crypto++ */
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55",
.klen = 8,
.iv = "\xE7\x82\x1D\xB8\x53\x11\xAC\x47",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x62\xE5\xF4\xDC\x99\xE7\x89\xE3"
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
"\xF4\x10\xCC\x21\x99\xEB\xDC\x15"
"\x19\x13\x93\x27\x9D\xB6\x6F\x45"
"\x17\x55\x61\x72\xC8\xD3\x7F\xA5"
"\x32\xD0\xD3\x02\x15\xA4\x05\x23"
"\x9C\x23\x61\x60\x77\x7B\x6C\x95"
"\x26\x49\x42\x2E\xF3\xC1\x8C\x6D"
"\xC8\x47\xD5\x94\xE7\x53\xC8\x23"
"\x1B\xA5\x0B\xCB\x12\xD3\x7A\x12"
"\xA4\x42\x15\x34\xF7\x5F\xDC\x58"
"\x5B\x58\x4C\xAD\xD1\x33\x8E\xE6"
"\xE5\xA0\xDA\x4D\x94\x3D\x63\xA8"
"\x02\x82\xBB\x16\xB8\xDC\xB5\x58"
"\xC3\x2D\x79\xE4\x25\x79\x43\xF9"
"\x6D\xD3\xCA\xC0\xE8\x12\xD4\x7E"
"\x04\x25\x79\xFD\x27\xFB\xC4\xEA"
"\x32\x94\x48\x92\xF3\x68\x1A\x7F"
"\x36\x33\x43\x79\xF7\xCA\xC2\x38"
"\xC0\x68\xD4\x53\xA9\xCC\x43\x0C"
"\x40\x57\x3E\xED\x00\x9F\x22\x6E"
"\x80\x99\x0B\xCC\x40\x63\x46\x8A"
"\xE8\xC4\x9B\x6D\x7A\x08\x6E\xA9"
"\x6F\x84\xBC\xB3\xF4\x95\x0B\x2D"
"\x6A\xBA\x37\x50\xC3\xCF\x9F\x7C"
"\x59\x5E\xDE\x0B\x30\xFA\x34\x8A"
"\xF8\xD1\xA2\xF8\x4E\xBD\x5D\x5E"
"\x7D\x71\x99\xE0\xF6\xE5\x7C\xE0"
"\x6D\xEE\x82\x89\x92\xD4\xF5\xD7"
"\xDF\x85\x2D\xE1\xB2\xD6\xAB\x94"
"\xA5\xA6\xE7\xB0\x51\x36\x52\x37"
"\x91\x45\x05\x3E\x58\xBF\x32",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 247,
crypto: testmgr - add new larger DES testvectors Most DES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:07 -06:00
.also_non_np = 1,
.np = 2,
.tap = { 247 - 8, 8 },
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des3_ede_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* These are from openssl */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\x55\x55\x55\x55\x55\x55\x55\x55"
"\xfe\xdc\xba\x98\x76\x54\x32\x10",
.klen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x73\x6f\x6d\x65\x64\x61\x74\x61",
.ctext = "\x18\xd7\x48\xe5\x63\x62\x05\x72",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x03\x52\x02\x07\x67\x20\x82\x17"
"\x86\x02\x87\x66\x59\x08\x21\x98"
"\x64\x05\x6a\xbd\xfe\xa9\x34\x57",
.klen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x73\x71\x75\x69\x67\x67\x6c\x65",
.ctext = "\xc0\x7d\x2a\x0f\xa5\x66\xfa\x30",
.len = 8,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x10\x46\x10\x34\x89\x98\x80\x20"
"\x91\x07\xd0\x15\x89\x19\x01\x01"
"\x19\x07\x92\x10\x98\x1a\x01\x01",
.klen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\xe1\xef\x62\xc3\x32\xfe\x82\x5b",
.len = 8,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\xF3\x9C\xD6\xF3\x9C\xB9\x5A\x67"
"\x00\x5A\x67\x00\x2D\xCE\xEB\x2D"
"\xCE\xEB\xB4\x51\x72\xB4\x51\x72",
.klen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\xEC\x77\xFB\x42\xD5\x59\x20"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x8B\x12\x86\x69\xF0\x5B\xCF\x56"
"\x39\xAD\x34\x9F\x66\xEA\x7D\xC4"
"\x48\xD3\xBA\x0D\xB1\x18\xE3\x4A"
"\xFE\x41\x28\x5C\x27\x8E\x11\x85"
"\x6C\xF7\x5E\xC2\x55\x3C\xA0\x0B"
"\x92\x65\xE9\x70\xDB\x4F\xD6\xB9"
"\x00\xB4\x1F\xE6\x49\xFD\x44\x2F"
"\x53\x3A\x8D\x14\x98\x63\xCA\x5D"
"\xC1\xA8\x33\xA7\x0E\x91\x78\xEC"
"\x77\xDE\x42\xD5\xBC\x07\x8B\x12"
"\xE5\x4C\xF0\x5B\x22\x56\x39\x80"
"\x6B\x9F\x66\xC9\x50\xC4\xAF\x36"
"\xBA\x0D\x94\x7F\xE3\x4A\xDD\x41"
"\x28\xB3\x1A\x8E\x11\xF8\x43\xF7"
"\x5E\x21\x55\x3C\x87\x6E\x92\x65"
"\xCC\x57\xDB\xA2\x35\xB9\x00\xEB"
"\x72\xE6\x49\xD0\x44\x2F\xB6\x19"
"\x8D\x14\xFF\x46\xCA\x5D\x24\xA8"
"\x33\x9A\x6D\x91\x78\xC3\x77\xDE"
"\xA1\x08\xBC\x07\xEE\x71\xE5\x4C"
"\xD7\x5B\x22\xB5\x1C\x80\x6B\xF2"
"\x45\xC9\x50\x3B\xAF\x36\x99\x60"
"\x94\x7F\xC6\x4A\xDD\xA4\x0F\xB3"
"\x1A\xED\x74\xF8\x43\x2A\x5E\x21"
"\x88\x13\x87\x6E\xF1\x58\xCC\x57"
"\x3E\xA2\x35\x9C\x67\xEB\x72\xC5"
"\x49\xD0\xBB\x02\xB6\x19\xE0\x4B"
"\xFF\x46\x29\x5D\x24\x8F\x16\x9A"
"\x6D\xF4\x5F\xC3\xAA\x3D\xA1\x08"
"\x93\x7A\xEE\x71\xD8\x4C\xD7\xBE"
"\x01\xB5\x1C\xE7\x4E\xF2\x45\x2C"
"\x50\x3B\x82\x15\x99\x60\xCB\x52"
"\xC6\xA9\x30\xA4\x0F\x96\x79\xED"
"\x74\xDF\x43\x2A\xBD\x04\x88\x13"
"\xFA\x4D\xF1\x58\x23\x57\x3E\x81"
"\x68\x9C\x67\xCE\x51\xC5\xAC\x37"
"\xBB\x02\x95\x7C\xE0\x4B\xD2\x46"
"\x29\xB0\x1B\x8F\x16\xF9\x40\xF4"
"\x5F\x26\xAA\x3D\x84\x6F\x93\x7A"
"\xCD\x54\xD8\xA3\x0A\xBE\x01\xE8"
"\x73\xE7\x4E\xD1\x45\x2C\xB7\x1E"
"\x82\x15\xFC\x47\xCB\x52\x25\xA9"
"\x30\x9B\x62\x96\x79\xC0\x74\xDF"
"\xA6\x09\xBD\x04\xEF\x76\xFA\x4D"
"\xD4\x58\x23\x8A\x1D\x81\x68\xF3"
"\x5A\xCE\x51\x38\xAC\x37\x9E\x61"
"\x95\x7C\xC7\x4B\xD2\xA5\x0C\xB0"
"\x1B\xE2\x75\xF9\x40\x2B\x5F\x26"
"\x89\x10\x84\x6F\xF6\x59\xCD\x54"
"\x3F\xA3\x0A\x9D\x64\xE8\x73\xDA"
"\x4E\xD1\xB8\x03\xB7\x1E\xE1\x48"
"\xFC\x47\x2E\x52\x25\x8C\x17\x9B"
"\x62\xF5\x5C\xC0\xAB\x32\xA6\x09"
"\x90\x7B\xEF\x76\xD9\x4D\xD4\xBF"
"\x06\x8A\x1D\xE4\x4F\xF3\x5A\x2D"
"\x51\x38\x83\x6A\x9E\x61\xC8\x53"
"\xC7\xAE\x31\xA5\x0C\x97\x7E\xE2"
"\x75\xDC\x40\x2B\xB2\x05\x89\x10"
"\xFB\x42\xF6\x59\x20\x54\x3F\x86"
"\x69\x9D\x64\xCF\x56\xDA\xAD\x34"
"\xB8\x03\xEA\x7D\xE1\x48\xD3\x47",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x4E\x9A\x40\x3D\x61\x7D\x17\xFA"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x16\x86\x88\x0B\xD8\xAE\xF8\xE4"
"\x81\x01\x04\x00\x76\xFA\xED\xD3"
"\x44\x7E\x21\x9D\xF0\xFB\x2B\x64"
"\xCA\x4E\x90\xE0\xC0\x63\x28\x92"
"\xF3\x1F\xA4\x53\x2C\x77\xCC\x77"
"\x69\x56\xD0\x19\xAD\x00\x2D\x97"
"\xBC\xDE\x49\x6A\x82\xBC\x16\xE2"
"\x2F\x3E\x72\xEE\xD1\xCE\xFC\x1B"
"\xEA\x32\x56\xE4\x0B\xAF\x27\x36"
"\xAF\x08\xB9\x61\xB7\x48\x23\x27"
"\xEE\x4D\xC8\x79\x56\x06\xEB\xC7"
"\x5B\xCA\x0A\xC6\x5E\x5C\xCB\xB6"
"\x9D\xDA\x04\x59\xE2\x09\x48\x7E"
"\x6B\x37\xC6\xFE\x92\xA9\x1E\x6E"
"\x0D\x19\xFA\x33\x0F\xEE\x36\x68"
"\x11\xBB\xF9\x5A\x73\xAB\x3A\xEA"
"\xAC\x28\xD8\xD5\x27\xE8\x6B\x16"
"\x45\x86\x50\x01\x70\x35\x99\x92"
"\xDF\x0C\x07\x88\x8B\x7F\x9E\x4B"
"\xD2\x04\x84\x90\xC4\x27\xDF\x0A"
"\x49\xA8\xA7\x1A\x6D\x78\x16\xCA"
"\xB3\x18\x5C\xC3\x93\x63\x5A\x68"
"\x77\x02\xBA\xED\x62\x71\xB1\xD9"
"\x5E\xE5\x6F\x1A\xCC\x1D\xBE\x2E"
"\x11\xF3\xA6\x97\xCA\x8E\xBF\xB4"
"\x56\xA1\x36\x6B\xB1\x0A\x3E\x70"
"\xEA\xD7\xCD\x72\x7B\x79\xC8\xAD"
"\x6B\xFE\xFB\xBA\x64\xAE\x19\xC1"
"\x82\xCF\x8A\xA1\x50\x17\x7F\xB2"
"\x6F\x7B\x0F\x52\xC5\x3E\x4A\x52"
"\x3F\xD9\x3F\x01\xA6\x41\x1A\xB3"
"\xB3\x7A\x0E\x8E\x75\xB2\xB1\x5F"
"\xDB\xEA\x84\x13\x26\x6C\x85\x4E"
"\xAE\x6B\xDC\xE7\xE7\xAD\xB0\x06"
"\x5C\xBA\x92\xD0\x30\xBB\x8D\xD2"
"\xAE\x4C\x70\x85\xA0\x07\xE3\x2C"
"\xD1\x27\x9C\xCF\xDB\x13\xB7\xE5"
"\xF9\x6A\x02\xD0\x39\x9D\xB6\xE7"
"\xD1\x17\x25\x08\xF9\xA9\xA6\x67"
"\x38\x80\xD1\x22\xAB\x1A\xD7\x26"
"\xAD\xCA\x19\x1B\xFA\x18\xA7\x57"
"\x31\xEC\xC9\xED\xDB\x79\xC0\x48"
"\xAC\x31\x9F\x03\x8B\x62\x5B\x7E"
"\x0E\xA6\xD0\x64\xEE\xEA\x00\xFC"
"\x58\xC8\xDE\x51\x4E\x17\x15\x11"
"\x66\x58\xB6\x90\xDC\xDF\xA1\x49"
"\xCA\x79\xE9\x31\x31\x42\xDC\x56"
"\x0B\xCD\xB6\x0D\xC7\x64\xF7\x19"
"\xD9\x42\x05\x7F\xBC\x2F\xFC\x90"
"\xAE\x29\x86\xAA\x43\x7A\x4F\x6B"
"\xCE\xEA\xBC\x31\x8D\x65\x9D\x46"
"\xEA\x77\xB4\xF9\x58\xEA\x5D\x84"
"\xE4\xDC\x14\xBB\xBD\x15\x0E\xDA"
"\xD8\xE4\xA4\x5D\x61\xF9\x58\x0F"
"\xE4\x82\x77\xCE\x87\xC0\x09\xF0"
"\xD6\x10\x9E\x34\xE1\x0C\x67\x55"
"\x7B\x6D\xD5\x51\x4B\x00\xEE\xBA"
"\xF2\x7B\xBE\x75\x07\x42\x9D\x99"
"\x12\xE1\x71\x4A\xF9\x2A\xF5\xF6"
"\x93\x03\xD7\x51\x09\xFA\xBE\x68"
"\xD8\x45\xFF\x33\xBA\xBB\x2B\x63",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 496,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des3_ede_cbc_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* Generated from openssl */
.key = "\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 128,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\x9C\xD6\xF3\x9C\xB9\x5A\x67\x00"
"\x5A\x67\x00\x2D\xCE\xEB\x2D\xCE"
"\xEB\xB4\x51\x72\xB4\x51\x72\x1F",
.klen = 24,
.iv = "\xB2\xD7\x48\xED\x06\x44\xF9\x12"
"\xB7\x28\x4D\x83\x24\x59\xF2\x17",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\xEC\x77\xFB\x42\xD5\x59\x20"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x8B\x12\x86\x69\xF0\x5B\xCF\x56"
"\x39\xAD\x34\x9F\x66\xEA\x7D\xC4"
"\x48\xD3\xBA\x0D\xB1\x18\xE3\x4A"
"\xFE\x41\x28\x5C\x27\x8E\x11\x85"
"\x6C\xF7\x5E\xC2\x55\x3C\xA0\x0B"
"\x92\x65\xE9\x70\xDB\x4F\xD6\xB9"
"\x00\xB4\x1F\xE6\x49\xFD\x44\x2F"
"\x53\x3A\x8D\x14\x98\x63\xCA\x5D"
"\xC1\xA8\x33\xA7\x0E\x91\x78\xEC"
"\x77\xDE\x42\xD5\xBC\x07\x8B\x12"
"\xE5\x4C\xF0\x5B\x22\x56\x39\x80"
"\x6B\x9F\x66\xC9\x50\xC4\xAF\x36"
"\xBA\x0D\x94\x7F\xE3\x4A\xDD\x41"
"\x28\xB3\x1A\x8E\x11\xF8\x43\xF7"
"\x5E\x21\x55\x3C\x87\x6E\x92\x65"
"\xCC\x57\xDB\xA2\x35\xB9\x00\xEB"
"\x72\xE6\x49\xD0\x44\x2F\xB6\x19"
"\x8D\x14\xFF\x46\xCA\x5D\x24\xA8"
"\x33\x9A\x6D\x91\x78\xC3\x77\xDE"
"\xA1\x08\xBC\x07\xEE\x71\xE5\x4C"
"\xD7\x5B\x22\xB5\x1C\x80\x6B\xF2"
"\x45\xC9\x50\x3B\xAF\x36\x99\x60"
"\x94\x7F\xC6\x4A\xDD\xA4\x0F\xB3"
"\x1A\xED\x74\xF8\x43\x2A\x5E\x21"
"\x88\x13\x87\x6E\xF1\x58\xCC\x57"
"\x3E\xA2\x35\x9C\x67\xEB\x72\xC5"
"\x49\xD0\xBB\x02\xB6\x19\xE0\x4B"
"\xFF\x46\x29\x5D\x24\x8F\x16\x9A"
"\x6D\xF4\x5F\xC3\xAA\x3D\xA1\x08"
"\x93\x7A\xEE\x71\xD8\x4C\xD7\xBE"
"\x01\xB5\x1C\xE7\x4E\xF2\x45\x2C"
"\x50\x3B\x82\x15\x99\x60\xCB\x52"
"\xC6\xA9\x30\xA4\x0F\x96\x79\xED"
"\x74\xDF\x43\x2A\xBD\x04\x88\x13"
"\xFA\x4D\xF1\x58\x23\x57\x3E\x81"
"\x68\x9C\x67\xCE\x51\xC5\xAC\x37"
"\xBB\x02\x95\x7C\xE0\x4B\xD2\x46"
"\x29\xB0\x1B\x8F\x16\xF9\x40\xF4"
"\x5F\x26\xAA\x3D\x84\x6F\x93\x7A"
"\xCD\x54\xD8\xA3\x0A\xBE\x01\xE8"
"\x73\xE7\x4E\xD1\x45\x2C\xB7\x1E"
"\x82\x15\xFC\x47\xCB\x52\x25\xA9"
"\x30\x9B\x62\x96\x79\xC0\x74\xDF"
"\xA6\x09\xBD\x04\xEF\x76\xFA\x4D"
"\xD4\x58\x23\x8A\x1D\x81\x68\xF3"
"\x5A\xCE\x51\x38\xAC\x37\x9E\x61"
"\x95\x7C\xC7\x4B\xD2\xA5\x0C\xB0"
"\x1B\xE2\x75\xF9\x40\x2B\x5F\x26"
"\x89\x10\x84\x6F\xF6\x59\xCD\x54"
"\x3F\xA3\x0A\x9D\x64\xE8\x73\xDA"
"\x4E\xD1\xB8\x03\xB7\x1E\xE1\x48"
"\xFC\x47\x2E\x52\x25\x8C\x17\x9B"
"\x62\xF5\x5C\xC0\xAB\x32\xA6\x09"
"\x90\x7B\xEF\x76\xD9\x4D\xD4\xBF"
"\x06\x8A\x1D\xE4\x4F\xF3\x5A\x2D"
"\x51\x38\x83\x6A\x9E\x61\xC8\x53"
"\xC7\xAE\x31\xA5\x0C\x97\x7E\xE2"
"\x75\xDC\x40\x2B\xB2\x05\x89\x10"
"\xFB\x42\xF6\x59\x20\x54\x3F\x86"
"\x69\x9D\x64\xCF\x56\xDA\xAD\x34"
"\xB8\x03\xEA\x7D\xE1\x48\xD3\x47",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xF8\xF6\xB5\x60\x5C\x5A\x75\x84"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x87\x81\x53\xBA\xC9\x6F\xEC\xD5"
"\x1E\x68\x8E\x85\x12\x86\x1D\x38"
"\x1C\x91\x40\xCC\x69\x6A\xD5\x35"
"\x0D\x7C\xB5\x07\x7C\x7B\x2A\xAF"
"\x32\xBC\xA1\xB3\x84\x31\x1B\x3C"
"\x0A\x2B\xFA\xD3\x9F\xB0\x8C\x37"
"\x8F\x9D\xA7\x6D\x6C\xFA\xD7\x90"
"\xE3\x69\x54\xED\x3A\xC4\xF1\x6B"
"\xB1\xCC\xFB\x7D\xD8\x8E\x17\x0B"
"\x9C\xF6\x4C\xD6\xFF\x03\x4E\xD9"
"\xE6\xA5\xAD\x25\xE6\x17\x69\x63"
"\x11\x35\x61\x94\x88\x7B\x1C\x48"
"\xF1\x24\x20\x29\x6B\x93\x1A\x8E"
"\x43\x03\x89\xD8\xB1\xDA\x47\x7B"
"\x79\x3A\x83\x76\xDA\xAE\xC6\xBB"
"\x22\xF8\xE8\x3D\x9A\x65\x54\xD8"
"\x4C\xE9\xE7\xE4\x63\x2F\x5C\x73"
"\x5A\xC3\xAE\x46\xA8\xCD\x57\xE6"
"\x67\x88\xA5\x20\x6F\x5F\x97\xC7"
"\xCC\x15\xA2\x0A\x93\xEA\x33\xE7"
"\x03\x5F\xEC\x64\x30\x6F\xEE\xD7"
"\x7E\xDF\xD6\xE9\x6F\x3F\xD6\x1E"
"\xBE\x67\x6C\x5B\x97\xA0\x09\xE6"
"\xEE\xFE\x55\xA3\x29\x65\xE0\x12"
"\xA1\x6A\x8A\x6F\xF2\xE6\xF1\x96"
"\x87\xFB\x9C\x05\xDD\x80\xEC\xFF"
"\xC5\xED\x50\xFE\xFC\x91\xCD\xCE"
"\x25\x2C\x5F\xD9\xAD\x95\x7D\x99"
"\xF0\x05\xC4\x71\x46\x5F\xF9\x0D"
"\xD2\x63\xDF\x9B\x96\x2E\x2B\xA6"
"\x2B\x1C\xD5\xFB\x96\x24\x60\x60"
"\x54\x40\xB8\x62\xA4\xF8\x46\x95"
"\x73\x28\xA3\xA6\x16\x2B\x17\xE7"
"\x7A\xF8\x62\x54\x3B\x64\x69\xE1"
"\x71\x34\x29\x5B\x4E\x05\x9B\xFA"
"\x5E\xF1\x96\xB7\xCE\x16\x9B\x59"
"\xF1\x1A\x4C\x51\x26\xFD\x79\xE2"
"\x3B\x8E\x71\x69\x6A\x91\xB6\x65"
"\x32\x09\xB8\xE4\x09\x1F\xEA\x39"
"\xCE\x20\x65\x9F\xD6\xD1\xC7\xF0"
"\x73\x50\x08\x56\x20\x9B\x94\x23"
"\x14\x39\xB7\x2B\xB1\x2D\x6D\x6F"
"\x41\x5B\xCC\xE2\x18\xAE\x62\x89"
"\x78\x8E\x67\x23\xD0\xFB\x2B\xE5"
"\x25\xC9\x48\x97\xB5\xD3\x17\xD5"
"\x6A\x9F\xA7\x48\x0C\x2B\x73\x3B"
"\x57\x08\xAE\x91\xF2\xB7\x57\x89"
"\xF4\xD0\xB0\x07\xB0\x42\x6C\xAF"
"\x98\x1A\xE7\xD1\xAC\x1E\xB5\x02"
"\xD4\x56\x42\x79\x79\x7F\x2A\x77"
"\x25\xE9\x7D\xC1\x88\x19\x2B\x49"
"\x6F\x46\x59\xAB\x56\x1F\x61\xE0"
"\x0C\x24\x9C\xC9\x5B\x63\xA9\x12"
"\xCF\x88\x96\xB6\xA8\x24\xC6\xA8"
"\x21\x85\x1A\x62\x7E\x34\xBB\xEB"
"\xBD\x02\x2A\xC7\xD8\x89\x80\xC5"
"\xB1\xBB\x60\xA5\x22\xFC\x6F\x38"
"\x02\x80\xA3\x28\x22\x75\xE1\xE9"
"\x90\xE9\xFA\x4B\x00\x10\xAC\x58"
"\x83\x70\xFF\x86\xE6\xAA\x0F\x1F"
"\x95\x63\x73\xA2\x44\xAC\xF8\xA5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 496,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec des3_ede_ctr_tv_template[] = {
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
{ /* Generated with Crypto++ */
.key = "\x9C\xD6\xF3\x9C\xB9\x5A\x67\x00"
"\x5A\x67\x00\x2D\xCE\xEB\x2D\xCE"
"\xEB\xB4\x51\x72\xB4\x51\x72\x1F",
.klen = 24,
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\xEC\x77\xFB\x42\xD5\x59\x20"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x8B\x12\x86\x69\xF0\x5B\xCF\x56"
"\x39\xAD\x34\x9F\x66\xEA\x7D\xC4"
"\x48\xD3\xBA\x0D\xB1\x18\xE3\x4A"
"\xFE\x41\x28\x5C\x27\x8E\x11\x85"
"\x6C\xF7\x5E\xC2\x55\x3C\xA0\x0B"
"\x92\x65\xE9\x70\xDB\x4F\xD6\xB9"
"\x00\xB4\x1F\xE6\x49\xFD\x44\x2F"
"\x53\x3A\x8D\x14\x98\x63\xCA\x5D"
"\xC1\xA8\x33\xA7\x0E\x91\x78\xEC"
"\x77\xDE\x42\xD5\xBC\x07\x8B\x12"
"\xE5\x4C\xF0\x5B\x22\x56\x39\x80"
"\x6B\x9F\x66\xC9\x50\xC4\xAF\x36"
"\xBA\x0D\x94\x7F\xE3\x4A\xDD\x41"
"\x28\xB3\x1A\x8E\x11\xF8\x43\xF7"
"\x5E\x21\x55\x3C\x87\x6E\x92\x65"
"\xCC\x57\xDB\xA2\x35\xB9\x00\xEB"
"\x72\xE6\x49\xD0\x44\x2F\xB6\x19"
"\x8D\x14\xFF\x46\xCA\x5D\x24\xA8"
"\x33\x9A\x6D\x91\x78\xC3\x77\xDE"
"\xA1\x08\xBC\x07\xEE\x71\xE5\x4C"
"\xD7\x5B\x22\xB5\x1C\x80\x6B\xF2"
"\x45\xC9\x50\x3B\xAF\x36\x99\x60"
"\x94\x7F\xC6\x4A\xDD\xA4\x0F\xB3"
"\x1A\xED\x74\xF8\x43\x2A\x5E\x21"
"\x88\x13\x87\x6E\xF1\x58\xCC\x57"
"\x3E\xA2\x35\x9C\x67\xEB\x72\xC5"
"\x49\xD0\xBB\x02\xB6\x19\xE0\x4B"
"\xFF\x46\x29\x5D\x24\x8F\x16\x9A"
"\x6D\xF4\x5F\xC3\xAA\x3D\xA1\x08"
"\x93\x7A\xEE\x71\xD8\x4C\xD7\xBE"
"\x01\xB5\x1C\xE7\x4E\xF2\x45\x2C"
"\x50\x3B\x82\x15\x99\x60\xCB\x52"
"\xC6\xA9\x30\xA4\x0F\x96\x79\xED"
"\x74\xDF\x43\x2A\xBD\x04\x88\x13"
"\xFA\x4D\xF1\x58\x23\x57\x3E\x81"
"\x68\x9C\x67\xCE\x51\xC5\xAC\x37"
"\xBB\x02\x95\x7C\xE0\x4B\xD2\x46"
"\x29\xB0\x1B\x8F\x16\xF9\x40\xF4"
"\x5F\x26\xAA\x3D\x84\x6F\x93\x7A"
"\xCD\x54\xD8\xA3\x0A\xBE\x01\xE8"
"\x73\xE7\x4E\xD1\x45\x2C\xB7\x1E"
"\x82\x15\xFC\x47\xCB\x52\x25\xA9"
"\x30\x9B\x62\x96\x79\xC0\x74\xDF"
"\xA6\x09\xBD\x04\xEF\x76\xFA\x4D"
"\xD4\x58\x23\x8A\x1D\x81\x68\xF3"
"\x5A\xCE\x51\x38\xAC\x37\x9E\x61"
"\x95\x7C\xC7\x4B\xD2\xA5\x0C\xB0"
"\x1B\xE2\x75\xF9\x40\x2B\x5F\x26"
"\x89\x10\x84\x6F\xF6\x59\xCD\x54"
"\x3F\xA3\x0A\x9D\x64\xE8\x73\xDA"
"\x4E\xD1\xB8\x03\xB7\x1E\xE1\x48"
"\xFC\x47\x2E\x52\x25\x8C\x17\x9B"
"\x62\xF5\x5C\xC0\xAB\x32\xA6\x09"
"\x90\x7B\xEF\x76\xD9\x4D\xD4\xBF"
"\x06\x8A\x1D\xE4\x4F\xF3\x5A\x2D"
"\x51\x38\x83\x6A\x9E\x61\xC8\x53"
"\xC7\xAE\x31\xA5\x0C\x97\x7E\xE2"
"\x75\xDC\x40\x2B\xB2\x05\x89\x10"
"\xFB\x42\xF6\x59\x20\x54\x3F\x86"
"\x69\x9D\x64\xCF\x56\xDA\xAD\x34"
"\xB8\x03\xEA\x7D\xE1\x48\xD3\x47",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x07\xC2\x08\x20\x72\x1F\x49\xEF"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x19\xCD\x6F\x32\x53\x05\x22\x15"
"\xA2\x85\x2B\xDB\x85\xD2\xD8\xB9"
"\xDD\x0D\x1B\x45\xCB\x69\x11\xD4"
"\xEA\xBE\xB2\x45\x5D\x0C\xAE\xBE"
"\xA0\xC1\x27\xAC\x65\x9F\x53\x7E"
"\xAF\xC2\x1B\xB5\xB8\x6D\x36\x0C"
"\x25\xC0\xF8\x6D\x0B\x29\x01\xDA"
"\x13\x78\xDC\x89\x12\x12\x43\xFA"
"\xF6\x12\xEF\x8D\x87\x62\x78\x83"
"\xE2\xBE\x41\x20\x4C\x6D\x35\x1B"
"\xD1\x0C\x30\xCF\xE2\xDE\x2B\x03"
"\xBF\x45\x73\xD4\xE5\x59\x95\xD1"
"\xB3\x9B\x27\x62\x97\xBD\xDE\x7F"
"\xA4\xD2\x39\x80\xAA\x50\x23\xF0"
"\x74\x88\x3D\xA8\x6A\x18\x79\x3B"
"\xC4\x96\x6C\x8D\x22\x40\x92\x6E"
"\xD6\xAD\x2A\x1F\xDE\x63\xC0\xE7"
"\x07\xF7\x2D\xF7\xB5\xF3\xF0\xCC"
"\x01\x7C\x2A\x9B\xC2\x10\xCA\xAA"
"\xFD\x2B\x3F\xC5\xF3\xF6\xFC\x9B"
"\x45\xDB\x53\xE4\x5B\xF3\xC9\x7B"
"\x8E\x52\xFF\xC8\x02\xB8\xAC\x9D"
"\xA1\x00\x39\xDA\x3D\x2D\x0E\x01"
"\x09\x7D\x8D\x5E\xBE\x53\xB9\xB0"
"\x8E\xE7\xE2\x96\x6A\xB2\x78\xEA"
"\xDE\x23\x8B\xA5\xFA\x5C\xE3\xDA"
"\xBF\x8E\x31\x6A\x55\xD1\x6A\xB2"
"\xB5\x46\x6F\xA5\xF0\xEE\xBA\x1F"
"\x9F\x98\xB0\x66\x4F\xD0\x3F\xA9"
"\xDF\x5F\x58\xC4\xF4\xFF\x75\x5C"
"\x40\x3A\x09\x7E\x6E\x1C\x97\xD4"
"\xCC\xE7\xE7\x71\xCF\x0B\x15\x08"
"\x71\xFA\x07\x97\xCD\xE6\xCA\x1D"
"\x14\x28\x0C\xCF\x99\x13\x7A\xF1"
"\xEB\xFA\xFA\x92\x07\xDE\x1D\xA1"
"\xD3\x36\x69\xFE\x51\x4D\x9F\x2E"
"\x83\x37\x4F\x1F\x48\x30\xED\x04"
"\x4D\xA4\xEF\x3A\xCA\x76\xF4\x1C"
"\x41\x8F\x63\x37\x78\x2F\x86\xA6"
"\xEF\x41\x7E\xD2\xAF\x88\xAB\x67"
"\x52\x71\xC3\x8E\xF8\x26\x93\x72"
"\xAA\xD6\x0E\xE7\x0B\x46\xB1\x3A"
"\xB4\x08\xA9\xA8\xA0\xCF\x20\x0C"
"\x52\xBC\x8B\x05\x56\xB2\xBC\x31"
"\x9B\x74\xB9\x29\x29\x96\x9A\x50"
"\xDC\x45\xDC\x1A\xEB\x0C\x64\xD4"
"\xD3\x05\x7E\x59\x55\xC3\xF4\x90"
"\xC2\xAB\xF8\x9B\x8A\xDA\xCE\xA1"
"\xC3\xF4\xAD\x77\xDD\x44\xC8\xAC"
"\xA3\xF1\xC9\xD2\x19\x5C\xB0\xCA"
"\xA2\x34\xC1\xF7\x6C\xFD\xAC\x65"
"\x32\xDC\x48\xC4\xF2\x00\x6B\x77"
"\xF1\x7D\x76\xAC\xC0\x31\x63\x2A"
"\xA5\x3A\x62\xC8\x91\xB1\x03\x65"
"\xCB\x43\xD1\x06\xDF\xC3\x67\xBC"
"\xDC\xE0\xCD\x35\xCE\x49\x65\xA0"
"\x52\x7B\xA7\x0D\x07\xA9\x1B\xB0"
"\x40\x77\x72\xC2\xEA\x0E\x3A\x78"
"\x46\xB9\x91\xB6\xE7\x3D\x51\x42"
"\xFD\x51\xB0\xC6\x2C\x63\x13\x78"
"\x5C\xEE\xFC\xCF\xC4\x70\x00\x34",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 496,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\x9C\xD6\xF3\x9C\xB9\x5A\x67\x00"
"\x5A\x67\x00\x2D\xCE\xEB\x2D\xCE"
"\xEB\xB4\x51\x72\xB4\x51\x72\x1F",
.klen = 24,
.iv = "\xB2\xD7\x48\xED\x06\x44\xF9\x12"
"\xB7\x28\x4D\x83\x24\x59\xF2\x17",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\xEC\x77\xFB\x42\xD5\x59\x20"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\x8B\x12\x86\x69\xF0\x5B\xCF\x56"
"\x39\xAD\x34\x9F\x66\xEA\x7D\xC4"
"\x48\xD3\xBA\x0D\xB1\x18\xE3\x4A"
"\xFE\x41\x28\x5C\x27\x8E\x11\x85"
"\x6C\xF7\x5E\xC2\x55\x3C\xA0\x0B"
"\x92\x65\xE9\x70\xDB\x4F\xD6\xB9"
"\x00\xB4\x1F\xE6\x49\xFD\x44\x2F"
"\x53\x3A\x8D\x14\x98\x63\xCA\x5D"
"\xC1\xA8\x33\xA7\x0E\x91\x78\xEC"
"\x77\xDE\x42\xD5\xBC\x07\x8B\x12"
"\xE5\x4C\xF0\x5B\x22\x56\x39\x80"
"\x6B\x9F\x66\xC9\x50\xC4\xAF\x36"
"\xBA\x0D\x94\x7F\xE3\x4A\xDD\x41"
"\x28\xB3\x1A\x8E\x11\xF8\x43\xF7"
"\x5E\x21\x55\x3C\x87\x6E\x92\x65"
"\xCC\x57\xDB\xA2\x35\xB9\x00\xEB"
"\x72\xE6\x49\xD0\x44\x2F\xB6\x19"
"\x8D\x14\xFF\x46\xCA\x5D\x24\xA8"
"\x33\x9A\x6D\x91\x78\xC3\x77\xDE"
"\xA1\x08\xBC\x07\xEE\x71\xE5\x4C"
"\xD7\x5B\x22\xB5\x1C\x80\x6B\xF2"
"\x45\xC9\x50\x3B\xAF\x36\x99\x60"
"\x94\x7F\xC6\x4A\xDD\xA4\x0F\xB3"
"\x1A\xED\x74\xF8\x43\x2A\x5E\x21"
"\x88\x13\x87\x6E\xF1\x58\xCC\x57"
"\x3E\xA2\x35\x9C\x67\xEB\x72\xC5"
"\x49\xD0\xBB\x02\xB6\x19\xE0\x4B"
"\xFF\x46\x29\x5D\x24\x8F\x16\x9A"
"\x6D\xF4\x5F\xC3\xAA\x3D\xA1\x08"
"\x93\x7A\xEE\x71\xD8\x4C\xD7\xBE"
"\x01\xB5\x1C\xE7\x4E\xF2\x45\x2C"
"\x50\x3B\x82\x15\x99\x60\xCB\x52"
"\xC6\xA9\x30\xA4\x0F\x96\x79\xED"
"\x74\xDF\x43\x2A\xBD\x04\x88\x13"
"\xFA\x4D\xF1\x58\x23\x57\x3E\x81"
"\x68\x9C\x67\xCE\x51\xC5\xAC\x37"
"\xBB\x02\x95\x7C\xE0\x4B\xD2\x46"
"\x29\xB0\x1B\x8F\x16\xF9\x40\xF4"
"\x5F\x26\xAA\x3D\x84\x6F\x93\x7A"
"\xCD\x54\xD8\xA3\x0A\xBE\x01\xE8"
"\x73\xE7\x4E\xD1\x45\x2C\xB7\x1E"
"\x82\x15\xFC\x47\xCB\x52\x25\xA9"
"\x30\x9B\x62\x96\x79\xC0\x74\xDF"
"\xA6\x09\xBD\x04\xEF\x76\xFA\x4D"
"\xD4\x58\x23\x8A\x1D\x81\x68\xF3"
"\x5A\xCE\x51\x38\xAC\x37\x9E\x61"
"\x95\x7C\xC7\x4B\xD2\xA5\x0C\xB0"
"\x1B\xE2\x75\xF9\x40\x2B\x5F\x26"
"\x89\x10\x84\x6F\xF6\x59\xCD\x54"
"\x3F\xA3\x0A\x9D\x64\xE8\x73\xDA"
"\x4E\xD1\xB8\x03\xB7\x1E\xE1\x48"
"\xFC\x47\x2E\x52\x25\x8C\x17\x9B"
"\x62\xF5\x5C\xC0\xAB\x32\xA6\x09"
"\x90\x7B\xEF\x76\xD9\x4D\xD4\xBF"
"\x06\x8A\x1D\xE4\x4F\xF3\x5A\x2D"
"\x51\x38\x83\x6A\x9E\x61\xC8\x53"
"\xC7\xAE\x31\xA5\x0C\x97\x7E\xE2"
"\x75\xDC\x40\x2B\xB2\x05\x89\x10"
"\xFB\x42\xF6\x59\x20\x54\x3F\x86"
"\x69\x9D\x64\xCF\x56\xDA\xAD\x34"
"\xB8\x03\xEA\x7D\xE1\x48\xD3\x47"
"\x2E\xB1\x18",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x23\xFF\x5C\x99\x75\xBB\x1F\xD4"
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
"\xBC\x27\x9D\x36\x60\xA9\xC9\xF7"
"\x94\x9D\x1B\xFF\x8E\x95\x57\x89"
"\x8C\x2E\x33\x70\x43\x61\xE6\xD2"
"\x82\x33\x63\xB6\xC4\x34\x5E\xF8"
"\x96\x07\xA7\xD2\x3B\x8E\xC9\xAA"
"\x7C\xA0\x55\x89\x2E\xE1\x85\x25"
"\x14\x04\xDA\x6B\xE0\xEE\x56\xCF"
"\x08\x2E\x69\xD4\x54\xDE\x22\x84"
"\x69\xA6\xA7\xD3\x3A\x9A\xE8\x05"
"\x63\xDB\xBF\x46\x3A\x26\x2E\x0F"
"\x58\x5C\x46\xEA\x07\x40\xDA\xE1"
"\x14\x1D\xCD\x4F\x06\xC0\xCA\x54"
"\x1E\xC9\x45\x85\x67\x7C\xC2\xB5"
"\x97\x5D\x61\x78\x2E\x46\xEC\x6A"
"\x53\xF4\xD0\xAE\xFA\xB4\x86\x29"
"\x9F\x17\x33\x24\xD8\xB9\xB2\x05"
"\x93\x88\xEA\xF7\xA0\x70\x69\x49"
"\x88\x6B\x73\x40\x41\x8D\xD9\xD9"
"\x7E\x78\xE9\xBE\x6C\x14\x22\x7A"
"\x66\xE1\xDA\xED\x10\xFF\x69\x1D"
"\xB9\xAA\xF2\x56\x72\x1B\x23\xE2"
"\x45\x54\x8B\xA3\x70\x23\xB4\x5E"
"\x8E\x96\xC9\x05\x00\xB3\xB6\xC2"
"\x2A\x02\x43\x7A\x62\xD5\xC8\xD2"
"\xC2\xD0\xE4\x78\xA1\x7B\x3E\xE8"
"\x9F\x7F\x7D\x40\x54\x30\x3B\xC0"
"\xA5\x54\xFD\xCA\x25\xEC\x44\x3E"
"\x1A\x54\x7F\x88\xD0\xE1\xFE\x71"
"\xCE\x05\x49\x89\xBA\xD6\x72\xE7"
"\xD6\x5D\x3F\xA2\xD9\xAB\xC5\x02"
"\xD6\x43\x22\xAF\xA2\xE4\x80\x85"
"\xD7\x87\xB9\xEA\x43\xDB\xC8\xEF"
"\x5C\x82\x2E\x98\x0D\x30\x41\x6B"
"\x08\x48\x8D\xF0\xF8\x60\xD7\x9D"
"\xE9\xDE\x40\xAD\x0D\xAD\x0D\x58"
"\x2A\x98\x35\xFE\xF7\xDD\x4B\x40"
"\xDE\xB0\x05\xD9\x7B\x09\x4D\xBC"
"\x42\xC0\xF1\x15\x0B\xFA\x26\x6B"
"\xC6\x12\x13\x4F\xCB\x35\xBA\x35"
"\xDD\x7A\x36\x9C\x12\x57\x55\x83"
"\x78\x58\x09\xD0\xB0\xCF\x7C\x5C"
"\x38\xCF\xBD\x79\x5B\x13\x4D\x97"
"\xC1\x85\x6F\x97\xC9\xE8\xC2\xA4"
"\x98\xE2\xBD\x77\x6B\x53\x39\x1A"
"\x28\x10\xE7\xE0\xE7\xDE\x9D\x69"
"\x78\x6F\x8E\xD2\xD9\x5D\xD2\x15"
"\x9E\xB5\x4D\x8C\xC0\x78\x22\x2F"
"\x17\x11\x2E\x99\xD7\xE3\xA4\x4F"
"\x65\xA5\x6B\x03\x2C\x35\x6F\xDA"
"\x8A\x19\x08\xE1\x08\x48\x59\x51"
"\x53\x4B\xD1\xDF\xDA\x14\x50\x5F"
"\xDF\xB5\x8C\xDF\xC6\xFD\x85\xFA"
"\xD4\xF9\x64\x45\x65\x0D\x7D\xF4"
"\xC8\xCD\x3F\x32\xAF\xDD\x30\xED"
"\x7B\xAA\xAC\xF0\xDA\x7F\xDF\x75"
"\x1C\xA4\xF1\xCB\x5E\x4F\x0B\xB4"
"\x97\x73\x28\xDE\xCF\xAF\x82\xBD"
"\xC4\xBA\xB4\x9C\x0D\x16\x77\x42"
"\x42\x39\x7C\x53\xA4\xD4\xDD\x40"
"\x5C\x60\x1F\x6E\xA7\xE2\xDC\xE7"
"\x32\x0F\x05\x2F\xF2\x4C\x95\x3B"
"\xF2\x79\xD9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 499,
crypto: testmgr - add new larger DES3_EDE testvectors Most DES3_EDE testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with DES3_EDE. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:12 -06:00
.also_non_np = 1,
.np = 2,
.tap = { 499 - 16, 16 },
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* Blowfish test vectors.
*/
static const struct cipher_testvec bf_tv_template[] = {
{ /* DES test vectors from OpenSSL */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 8,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x4e\xf9\x97\x45\x61\x98\xdd\x78",
.len = 8,
}, {
.key = "\x1f\x1f\x1f\x1f\x0e\x0e\x0e\x0e",
.klen = 8,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ctext = "\xa7\x90\x79\x51\x08\xea\x3c\xae",
.len = 8,
}, {
.key = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87",
.klen = 8,
.ptext = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\xe8\x7a\x24\x4e\x2c\xc8\x5e\x82",
.len = 8,
}, { /* Vary the keylength... */
.key = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87"
"\x78\x69\x5a\x4b\x3c\x2d\x1e\x0f",
.klen = 16,
.ptext = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\x93\x14\x28\x87\xee\x3b\xe1\x5c",
.len = 8,
}, {
.key = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87"
"\x78\x69\x5a\x4b\x3c\x2d\x1e\x0f"
"\x00\x11\x22\x33\x44",
.klen = 21,
.ptext = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\xe6\xf5\x1e\xd7\x9b\x9d\xb2\x1f",
.len = 8,
}, { /* Generated with bf488 */
.key = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87"
"\x78\x69\x5a\x4b\x3c\x2d\x1e\x0f"
"\x00\x11\x22\x33\x44\x55\x66\x77"
"\x04\x68\x91\x04\xc2\xfd\x3b\x2f"
"\x58\x40\x23\x64\x1a\xba\x61\x76"
"\x1f\x1f\x1f\x1f\x0e\x0e\x0e\x0e"
"\xff\xff\xff\xff\xff\xff\xff\xff",
.klen = 56,
.ptext = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\xc0\x45\x04\x01\x2e\x4e\x1f\x53",
.len = 8,
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x96\x87\x3D\x0C\x7B\xFB\xBD\x1F"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xE3\xC1\x99\x6D\x39\xD4\xC2\x7D"
"\xD7\x87\xA1\xF2\xDF\x51\x71\x26"
"\xC2\xF4\x6D\xFF\xF6\xCD\x6B\x40"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\xE1\xB3\xBF\xD4\x38\x2B\xC8\x3B"
"\xD3\xB2\xD4\x61\xC7\x9F\x06\xE9"
"\xCD\xF3\x88\x39\x39\x7A\xDF\x19"
"\xE8\x03\x2A\x0B\x9E\xA0\x2B\x86"
"\x31\xF8\x9D\xB1\xEE\x78\x9D\xB5"
"\xCD\x8B\x7C\x2E\xF5\xA2\x2D\x5D"
"\x6E\x66\xAF\x38\x6C\xD3\x13\xED"
"\x14\xEA\x5D\xD0\x17\x77\x0F\x4A"
"\x50\xF2\xD0\x0F\xC8\xF7\x1E\x7B"
"\x9D\x5B\x54\x65\x4F\x16\x8A\x97"
"\xF3\xF6\xD4\xAA\x87\x36\x77\x72"
"\x99\x4A\xB5\x5E\x88\xC3\xCD\x7D"
"\x1D\x97\xF9\x11\xBD\xE0\x1F\x1F"
"\x96\x3E\x4B\x22\xF4\xC0\xE6\xB8"
"\x47\x82\x98\x23\x33\x36\xBC\x1B"
"\x36\xE7\xF6\xCF\x97\x37\x16\xC0"
"\x87\x31\x8B\xB0\xDB\x19\x42\xA5"
"\x1F\x90\x7E\x66\x34\xDD\x5E\xE9"
"\x4F\xB2\x2B\x9A\xDE\xB3\x5D\x71"
"\x4D\x68\xF0\xDC\xA6\xEA\xE3\x9B"
"\x60\x00\x55\x57\x06\x8B\xD5\xB3"
"\x86\x30\x78\xDA\x33\x9A\x9D\xCC"
"\xBA\x0B\x81\x06\x77\x43\xC7\xC9"
"\xDB\x37\x60\x11\x45\x59\x6D\x2D"
"\x90\x3D\x65\x3E\xD0\x13\xC6\x3C"
"\x0E\x78\x7D\x9A\x00\xD6\x2F\x0B"
"\x3B\x53\x19\x1E\xA8\x9B\x11\xD9"
"\x98\xE4\x7F\xC3\x6E\x51\x24\x70"
"\x9F\x04\x9C\xC2\x9E\x44\x84\xE3"
"\xE0\x8A\x44\xA2\x5C\x94\x74\x34"
"\x37\x52\x7C\x03\xE8\x8E\x97\xE1"
"\x5B\x5C\x0E\xB0\x70\xFE\x54\x3F"
"\xD8\x65\xA9\xC5\xCD\xEC\xF4\x45"
"\x55\xC5\xA7\xA3\x19\x80\x28\x51"
"\xBE\x64\x4A\xC1\xD4\xE1\xBE\xEB"
"\x73\x4C\xB6\xF9\x5F\x6D\x82\xBC"
"\x3E\x42\x14\x49\x88\x51\xBF\x68"
"\x45\x75\x27\x1B\x0A\x72\xED\xAF"
"\xDA\xC4\x4D\x67\x0D\xEE\x75\xE3"
"\x34\xDD\x91\x19\x42\x3A\xCB\xDA"
"\x38\xFA\x3C\x93\x62\xF2\xE3\x81"
"\xB3\xE4\xBB\xF6\x0D\x0B\x1D\x09"
"\x9C\x52\x0D\x50\x63\xA4\xB2\xD2"
"\x82\xA0\x23\x3F\x1F\xB6\xED\x6E"
"\xC2\x9C\x1C\xD0\x9A\x40\xB6\xFC"
"\x36\x56\x6E\x85\x73\xD7\x52\xBA"
"\x35\x5E\x32\x89\x5D\x42\xF5\x36"
"\x52\x8D\x46\x7D\xC8\x71\xAD\x33"
"\xE1\xAF\x6A\xA8\xEC\xBA\x1C\xDC"
"\xFE\x88\xE6\x16\xE4\xC8\x13\x00"
"\x3C\xDA\x59\x32\x38\x19\xD5\xEB"
"\xB6\x7F\x78\x45\x1B\x8E\x07\x8C"
"\x66\x52\x75\xFF\xAF\xCE\x2D\x2B"
"\x22\x29\xCA\xB3\x5F\x7F\xE3\x29"
"\xB2\xB8\x9D\xEB\x16\xC8\xC5\x1D"
"\xC9\x0D\x59\x82\x27\x57\x9D\x42"
"\x54\x59\x09\xA5\x3D\xC5\x84\x68"
"\x56\xEB\x36\x77\x3D\xAA\xB8\xF5"
"\xC9\x1A\xFB\x5D\xDE\xBB\x43\xF4",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 504,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 504 - 10, 2, 8 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec bf_cbc_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* From OpenSSL */
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87",
.klen = 16,
.iv = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x37\x36\x35\x34\x33\x32\x31\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x4e\x6f\x77\x20\x69\x73\x20\x74"
"\x68\x65\x20\x74\x69\x6d\x65\x20"
"\x66\x6f\x72\x20\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x6b\x77\xb4\xd6\x30\x06\xde\xe6"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x05\xb1\x56\xe2\x74\x03\x97\x93"
"\x58\xde\xb9\xe7\x15\x46\x16\xd9"
"\x59\xf1\x65\x2b\xd5\xff\x92\xcc",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xB4\xFE\xA5\xBB\x3D\x2C\x27\x06"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\x06\x2B\x3A\x92\xB2\xF5\x5E\x62"
"\x84\xCD\xF7\x66\x7E\x41\x6C\x8E"
"\x1B\xD9\x02\xB6\x48\xB0\x87\x25"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x01\x9C\x93\x63\x51\x60\x82\xD2"
"\x4D\xE5\xC2\xB7\xAE\x60\xD8\xAD"
"\x9F\xAB\x6C\xFA\x20\x05\xDA\x6F"
"\x1F\xD1\xD8\x36\x0F\xB5\x16\x69"
"\x3C\xAF\xB3\x30\x18\x33\xE6\xB5"
"\x43\x29\x9D\x94\xF4\x2F\x0A\x65"
"\x40\xB2\xB2\xB2\x42\x89\xEE\x8A"
"\x60\xD3\x52\xA8\xED\x91\xDF\xE1"
"\x91\x73\x7C\x28\xA1\x14\xC3\x4C"
"\x82\x72\x4B\x7D\x7D\x32\xD5\x19"
"\xE8\xB8\x6B\x30\x21\x09\x0E\x27"
"\x10\x9D\x2D\x3A\x6A\x4B\x7B\xE6"
"\x8D\x4E\x02\x32\xFF\x7F\x8E\x13"
"\xB0\x96\xF4\xC2\xA1\x60\x8A\x69"
"\xEF\x0F\x86\xD0\x25\x13\x1A\x7C"
"\x6E\xF0\x41\xA3\xFB\xB3\xAB\x40"
"\x7D\x19\xA0\x11\x4F\x3E\x1D\x43"
"\x65\xFE\x15\x40\xD0\x62\x41\x02"
"\xEA\x0C\x7A\xC3\x84\xEE\xB0\xBE"
"\xBE\xC8\x57\x51\xCD\x4F\xAD\x5C"
"\xCC\x79\xBA\x0D\x85\x3A\xED\x6B"
"\xAC\x6B\xA3\x4D\xBC\xE8\x02\x6A"
"\xC2\x6D\xBD\x5E\x89\x95\x86\x43"
"\x2C\x17\x4B\xC6\x40\xA2\xBD\x24"
"\x04\xF0\x86\x08\x78\x18\x42\xE0"
"\x39\x1B\x22\x9E\x89\x4C\x04\x6B"
"\x65\xC5\xB6\x0E\xF6\x63\xFC\xD7"
"\xAE\x9E\x87\x13\xCC\xD3\x1A\xEC"
"\xF0\x51\xCC\x93\x68\xFC\xE9\x19"
"\x7C\x4E\x9B\xCC\x17\xAD\xD2\xFC"
"\x97\x18\x92\xFF\x15\x11\xCE\xED"
"\x04\x41\x05\xA3\x92\xFF\x3B\xE6"
"\xB6\x8C\x90\xC6\xCD\x15\xA0\x04"
"\x25\x8B\x5D\x5B\x5F\xDB\xAE\x68"
"\xEF\xB3\x61\x18\xDB\x83\x9B\x39"
"\xCA\x82\xD1\x88\xF0\xA2\x5C\x02"
"\x87\xBD\x8D\x8F\xBB\x62\xF0\x35"
"\x75\x6F\x06\x81\x0A\x97\x4D\xF0"
"\x43\x12\x73\x77\xDB\x91\x83\x5B"
"\xE7\x3A\xA6\x07\x7B\xBF\x2C\x50"
"\x94\xDE\x7B\x65\xDA\x1C\xF1\x9F"
"\x7E\x12\x40\xB2\x3E\x19\x23\xF1"
"\x7C\x1B\x5F\xA8\xF3\xAC\x63\x87"
"\xEB\x3E\x0C\xBE\xA3\x63\x97\x88"
"\x8D\x27\xC6\x2A\xF8\xF2\x67\x9A"
"\x0D\x14\x16\x2B\x6F\xCB\xD4\x76"
"\x14\x48\x2E\xDE\x2A\x44\x5E\x45"
"\xF1\x97\x82\xEF\xB7\xAE\xED\x3A"
"\xED\x73\xD3\x79\xF7\x38\x1D\xD0"
"\xC5\xF8\x69\x83\x28\x84\x87\x56"
"\x3F\xAE\x81\x04\x79\x1F\xD1\x09"
"\xC5\xE5\x05\x0D\x64\x16\xCE\x42"
"\xC5\xF8\xDB\x57\x89\x33\x22\xFC"
"\xB4\xD7\x94\xB9\xF3\xCC\x02\x90"
"\x02\xBA\x55\x1E\x24\x3E\x02\x1D"
"\xC6\xCD\x8F\xD9\xBD\xED\xB0\x51"
"\xCD\xE9\xD5\x0C\xFE\x12\x39\xA9"
"\x93\x9B\xEE\xB5\x97\x41\xD2\xA0"
"\xB4\x98\xD8\x6B\x74\xE7\x65\xF4",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 504,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 504 - 10, 2, 8 },
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec bf_ctr_tv_template[] = {
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
{ /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xC7\xA3\xDF\xB9\x05\xF4\x9E\x8D"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\x9E\xDF\x38\x18\x83\x07\xEF\xC1"
"\x93\x3C\xAA\xAA\xFE\x06\x42\xCC"
"\x0D\x70\x86\x5A\x44\xAD\x85\x17"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\xE4\x1F\x5E\xA5\x89\xAC\x32\xBC"
"\x3D\xA7\xE9\x0A\x5C\x70\x4D\xDE"
"\x99\x38\x07\xCA\x1D\x21\xC1\x11"
"\x97\xEB\x98\x75\xC4\x73\x45\x83"
"\x46\x1C\x9C\x91\x87\xC1\xA0\x56"
"\x98\xA1\x8B\xDB\x22\x76\xBD\x62"
"\xA4\xBC\xE8\x86\xDA\xD2\x51\x13"
"\x13\xD2\x96\x68\x69\x10\x67\x0C"
"\xD0\x17\x25\x7C\xB2\xAE\x4F\x93"
"\xA6\x82\x20\xCF\x0F\xA6\x47\x79"
"\x88\x09\x40\x59\xBD\x12\x64\xB5"
"\x19\x38\x0D\xFF\x86\xD9\x42\x20"
"\x81\x0D\x96\x99\xAF\x22\x1F\x94"
"\x5C\x6E\xEC\xEA\xA3\x39\xCB\x09"
"\x43\x19\x7F\xD0\xBB\x10\xC2\x49"
"\xF7\xE9\xF2\xEE\xBF\xF7\xF8\xB3"
"\x0E\x1A\xF1\x8D\x70\x82\x0C\x04"
"\xFD\x29\x1A\xAC\xC0\x92\x48\x34"
"\x6A\xE3\x1D\x4F\xFC\x1C\x72\x6A"
"\x57\xCB\xAD\xD0\x98\xAB\xB1\x01"
"\x03\x6A\x45\xDD\x07\x71\x5F\x5B"
"\xB5\x4A\xE4\xE5\xB9\xB9\xBC\xAC"
"\x44\xF7\x41\xA4\x5F\x2E\xE9\x28"
"\xE3\x05\xD2\x94\x78\x4C\x33\x1B"
"\xBD\xC1\x6E\x51\xD9\xAD\xD9\x86"
"\x15\x4A\x78\xAE\x7B\xAD\x3B\xBC"
"\x2F\xE0\x0E\xC5\x7B\x54\x97\x5F"
"\x60\x51\x14\x65\xF9\x91\xE9\xDA"
"\x9A\xBC\xFC\x19\x29\x67\xAA\x63"
"\x5E\xF2\x48\x88\xEB\x79\xE1\xE4"
"\xF7\xF6\x4C\xA9\xE2\x8C\x3B\xE0"
"\xED\x52\xAE\x90\x8F\x5B\x98\x34"
"\x29\x94\x34\x7F\xF9\x6C\x1E\xB6"
"\xA4\xE7\x2D\x06\x54\x9D\xC3\x02"
"\xC1\x90\xA4\x72\x31\x6B\x24\x51"
"\x0B\xB3\x7C\x63\x15\xBA\xAF\x5D"
"\x41\xE0\x37\x6D\xBE\x41\x58\xDE"
"\xF2\x07\x62\x99\xBE\xC1\x8C\x0F"
"\x0F\x28\xFB\x8F\x0E\x1D\x91\xE2"
"\xDA\x99\x5C\x49\xBA\x9C\xA8\x86"
"\x82\x63\x11\xB3\x54\x49\x00\x08"
"\x07\xF2\xE8\x1F\x34\x49\x61\xF4"
"\x81\xE9\xF6\xA9\x5A\x28\x60\x1F"
"\x66\x99\x08\x06\xF2\xE8\x2D\xD1"
"\xD0\x67\xBA\x32\x1F\x02\x86\x7B"
"\xFB\x79\x3D\xC5\xB1\x7F\x15\xAF"
"\xD7\xBF\x31\x46\x22\x7F\xAE\x5B"
"\x8B\x95\x47\xC2\xB1\x62\xA1\xCE"
"\x52\xAC\x9C\x8B\xC2\x49\x7F\xBC"
"\x9C\x89\xB8\xB6\xCA\xE3\x8F\xEA"
"\xAC\xB4\x5D\xE4\x50\xDC\x3A\xB5"
"\x91\x04\x94\x99\x03\x3B\x42\x6D"
"\x9C\x4A\x02\xF5\xB5\x38\x98\xA8"
"\x5C\x97\x2E\x4D\x79\x67\x71\xAF"
"\xF0\x70\x77\xFF\x2D\xDA\xA0\x9E"
"\x23\x8D\xD6\xA6\x68\x10\x78\x9A"
"\x64\xBB\x15\xB8\x56\xCF\xEE\xE5"
"\x32\x44\x96\x1C\xD8\xEB\x95\xD2"
"\xF3\x71\xEF\xEB\x4E\xBB\x4D\x29",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 504,
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xC7\xA3\xDF\xB9\x05\xF4\x9E\x8D"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\x9E\xDF\x38\x18\x83\x07\xEF\xC1"
"\x93\x3C\xAA\xAA\xFE\x06\x42\xCC"
"\x0D\x70\x86\x5A\x44\xAD\x85\x17"
"\xE4\x1F\x5E\xA5\x89\xAC\x32\xBC"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x3D\xA7\xE9\x0A\x5C\x70\x4D\xDE"
"\x99\x38\x07\xCA\x1D\x21\xC1\x11"
"\x97\xEB\x98\x75\xC4\x73\x45\x83"
"\x46\x1C\x9C\x91\x87\xC1\xA0\x56"
"\x98\xA1\x8B\xDB\x22\x76\xBD\x62"
"\xA4\xBC\xE8\x86\xDA\xD2\x51\x13"
"\x13\xD2\x96\x68\x69\x10\x67\x0C"
"\xD0\x17\x25\x7C\xB2\xAE\x4F\x93"
"\xA6\x82\x20\xCF\x0F\xA6\x47\x79"
"\x88\x09\x40\x59\xBD\x12\x64\xB5"
"\x19\x38\x0D\xFF\x86\xD9\x42\x20"
"\x81\x0D\x96\x99\xAF\x22\x1F\x94"
"\x5C\x6E\xEC\xEA\xA3\x39\xCB\x09"
"\x43\x19\x7F\xD0\xBB\x10\xC2\x49"
"\xF7\xE9\xF2\xEE\xBF\xF7\xF8\xB3"
"\x0E\x1A\xF1\x8D\x70\x82\x0C\x04"
"\xFD\x29\x1A\xAC\xC0\x92\x48\x34"
"\x6A\xE3\x1D\x4F\xFC\x1C\x72\x6A"
"\x57\xCB\xAD\xD0\x98\xAB\xB1\x01"
"\x03\x6A\x45\xDD\x07\x71\x5F\x5B"
"\xB5\x4A\xE4\xE5\xB9\xB9\xBC\xAC"
"\x44\xF7\x41\xA4\x5F\x2E\xE9\x28"
"\xE3\x05\xD2\x94\x78\x4C\x33\x1B"
"\xBD\xC1\x6E\x51\xD9\xAD\xD9\x86"
"\x15\x4A\x78\xAE\x7B\xAD\x3B\xBC"
"\x2F\xE0\x0E\xC5\x7B\x54\x97\x5F"
"\x60\x51\x14\x65\xF9\x91\xE9\xDA"
"\x9A\xBC\xFC\x19\x29\x67\xAA\x63"
"\x5E\xF2\x48\x88\xEB\x79\xE1\xE4"
"\xF7\xF6\x4C\xA9\xE2\x8C\x3B\xE0"
"\xED\x52\xAE\x90\x8F\x5B\x98\x34"
"\x29\x94\x34\x7F\xF9\x6C\x1E\xB6"
"\xA4\xE7\x2D\x06\x54\x9D\xC3\x02"
"\xC1\x90\xA4\x72\x31\x6B\x24\x51"
"\x0B\xB3\x7C\x63\x15\xBA\xAF\x5D"
"\x41\xE0\x37\x6D\xBE\x41\x58\xDE"
"\xF2\x07\x62\x99\xBE\xC1\x8C\x0F"
"\x0F\x28\xFB\x8F\x0E\x1D\x91\xE2"
"\xDA\x99\x5C\x49\xBA\x9C\xA8\x86"
"\x82\x63\x11\xB3\x54\x49\x00\x08"
"\x07\xF2\xE8\x1F\x34\x49\x61\xF4"
"\x81\xE9\xF6\xA9\x5A\x28\x60\x1F"
"\x66\x99\x08\x06\xF2\xE8\x2D\xD1"
"\xD0\x67\xBA\x32\x1F\x02\x86\x7B"
"\xFB\x79\x3D\xC5\xB1\x7F\x15\xAF"
"\xD7\xBF\x31\x46\x22\x7F\xAE\x5B"
"\x8B\x95\x47\xC2\xB1\x62\xA1\xCE"
"\x52\xAC\x9C\x8B\xC2\x49\x7F\xBC"
"\x9C\x89\xB8\xB6\xCA\xE3\x8F\xEA"
"\xAC\xB4\x5D\xE4\x50\xDC\x3A\xB5"
"\x91\x04\x94\x99\x03\x3B\x42\x6D"
"\x9C\x4A\x02\xF5\xB5\x38\x98\xA8"
"\x5C\x97\x2E\x4D\x79\x67\x71\xAF"
"\xF0\x70\x77\xFF\x2D\xDA\xA0\x9E"
"\x23\x8D\xD6\xA6\x68\x10\x78\x9A"
"\x64\xBB\x15\xB8\x56\xCF\xEE\xE5"
"\x32\x44\x96\x1C\xD8\xEB\x95\xD2"
"\xF3\x71\xEF\xEB\x4E\xBB\x4D",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 503,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
.np = 2,
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
.tap = { 503 - 8, 8 },
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x5F\x58\x6E\x60\x51\x6E\xDC\x3D"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xD1\xBB\xF7\xB7\xFD\x04\x44\x82"
"\xDC\x9F\x4B\x02\xF1\xD2\x5A\x6F"
"\x25\xF9\x27\x21\xF2\xD2\x9A\x01"
"\xBD\xAD\x3D\x93\x87\xCA\x0D\xFE"
"\xB7\x2C\x17\x1F\x42\x8C\x13\xB2"
"\x62\x44\x72\xB9\x5D\xC0\xF8\x37"
"\xDF\xEA\x78\x81\x8F\xA6\x34\xB2"
"\x07\x09\x7C\xB9\x3A\xA0\x2B\x18"
"\x34\x6A\x9D\x3D\xA5\xEB\xF4\x60"
"\xF8\x98\xA2\x39\x81\x23\x6C\xA9"
"\x70\xCA\xCC\x45\xD8\x1F\xDF\x44"
"\x2A\x67\x7A\x88\x28\xDC\x36\x83"
"\x18\xD7\x48\x43\x17\x2B\x1B\xE6"
"\x0B\x82\x59\x14\x26\x67\x08\x09"
"\x5B\x5D\x38\xD0\x81\xCE\x54\x2A"
"\xCD\x22\x94\x42\xF5\xBA\x74\x7E"
"\xD9\x00\x40\xA9\x0D\x0B\xBD\x8E"
"\xC4\x8E\x5E\x17\x8F\x48\xE2\xB8"
"\xF4\xCC\x19\x76\xAB\x48\x29\xAA"
"\x81\xD5\xCE\xD5\x8A\x3B\xC9\x21"
"\xEF\x50\x4F\x04\x02\xBF\xE1\x1F"
"\x59\x28\x1A\xE4\x18\x16\xA0\x29"
"\xBF\x34\xA9\x2D\x28\x83\xC0\x5E"
"\xEA\x44\xC4\x6E\xAB\x24\x79\x9D"
"\x2D\xA1\xE8\x55\xCA\x74\xFC\xBD"
"\xFE\xDD\xDA\xA5\xFB\x34\x90\x31"
"\x0E\x62\x28\x9B\xDC\xD7\xA1\xBB"
"\xF0\x1A\xB3\xE2\xD0\xFA\xBD\xE8"
"\x5C\x5A\x10\x67\xF6\x6A\x17\x3F"
"\xC5\xE9\x09\x08\xDD\x22\x77\x42"
"\x26\x6A\x6A\x7A\x3F\x87\x80\x0C"
"\xF0\xFF\x15\x8E\x84\x86\xC0\x10"
"\x0F\x8D\x33\x06\xB8\x72\xA4\x47"
"\x6B\xED\x2E\x05\x94\x6C\x5C\x5B"
"\x13\xF6\x77\xEE\x3B\x16\xDF\xC2"
"\x63\x66\x07\x6D\x3F\x6C\x51\x7C"
"\x1C\xAC\x80\xB6\x58\x48\xB7\x9D"
"\xB4\x19\xD8\x19\x45\x66\x27\x02"
"\xA1\xA9\x99\xF3\x1F\xE5\xA7\x1D"
"\x31\xE7\x1B\x0D\xFF\xBB\xB5\xA1"
"\xF5\x9C\x45\x1E\x18\x19\xA1\xE7"
"\xC2\xF1\xBF\x68\xC3\xEC\xCF\x53"
"\x67\xA6\x2B\x7D\x3C\x6D\x24\xC3"
"\xE8\xE6\x07\x5A\x09\xE0\x32\xA8"
"\x52\xF6\xE9\xED\x0E\xC6\x0A\x6A"
"\xFC\x60\x2A\xE0\x93\xCE\xB8\x2E"
"\xA2\xA8\x0E\x79\x9E\x34\x5D\x37"
"\x6F\x12\xFE\x48\x7B\xE7\xB9\x22"
"\x29\xE8\xD7\xBE\x5D\xD1\x8B\xD9"
"\x91\x51\x4E\x71\xF2\x98\x85\x16"
"\x25\x7A\x76\x8A\x51\x0E\x65\x14"
"\x81\xB5\x3A\x37\xFD\xEC\xB5\x8A"
"\xE1\xCF\x41\x72\x14\x29\x4C\xF0"
"\x20\xD9\x9A\xC5\x66\xA4\x03\x76"
"\x5B\xA4\x15\x4F\x0E\x64\x39\x40"
"\x25\xF9\x20\x22\xF5\x88\xF5\xBA"
"\xE4\xDF\x45\x61\xBF\x8D\x7A\x24"
"\x4B\x92\x71\xD9\x2F\x77\xA7\x95"
"\xA8\x7F\x61\xD5\xA4\x57\xB0\xFB"
"\xB5\x77\xBA\x1C\xEE\x71\xFA\xB0"
"\x16\x4C\x18\x6B\xF2\x69\xA0\x07"
"\xEF\xBE\xEC\x69\xAC\xA8\x63\x9E",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 504,
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* Twofish test vectors.
*/
static const struct cipher_testvec tf_tv_template[] = {
{
.key = zeroed_string,
.klen = 16,
.ptext = zeroed_string,
.ctext = "\x9f\x58\x9f\x5c\xf6\x12\x2c\x32"
"\xb6\xbf\xec\x2f\x2a\xe8\xc3\x5a",
.len = 16,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10"
"\x00\x11\x22\x33\x44\x55\x66\x77",
.klen = 24,
.ptext = zeroed_string,
.ctext = "\xcf\xd1\xd2\xe5\xa9\xbe\x9c\xdf"
"\x50\x1f\x13\xb8\x92\xbd\x22\x48",
.len = 16,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10"
"\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff",
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = zeroed_string,
.ctext = "\x37\x52\x7b\xe0\x05\x23\x34\xb8"
"\x9f\x0c\xfc\xca\xe8\x7c\xfa\x20",
.len = 16,
}, { /* Generated with Crypto++ */
.key = "\x3F\x85\x62\x3F\x1C\xF9\xD6\x1C"
"\xF9\xD6\xB3\x90\x6D\x4A\x90\x6D"
"\x4A\x27\x04\xE1\x27\x04\xE1\xBE"
"\x9B\x78\xBE\x9B\x78\x55\x32\x0F",
.klen = 32,
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\x88\xCB\x1E\xC2\xAF\x8A\x97\xFF"
"\xF6\x90\x46\x9C\x4A\x0F\x08\xDC"
"\xDE\xAB\xAD\xFA\xFC\xA8\xC2\x3D"
"\xE0\xE4\x8B\x3F\xD5\xA3\xF7\x14"
"\x34\x9E\xB6\x08\xB2\xDD\xA8\xF5"
"\xDF\xFA\xC7\xE8\x09\x50\x76\x08"
"\xA2\xB6\x6A\x59\xC0\x2B\x6D\x05"
"\x89\xF6\x82\xF0\xD3\xDB\x06\x02"
"\xB5\x11\x5C\x5E\x79\x1A\xAC\x43"
"\x5C\xC0\x30\x4B\x6B\x16\xA1\x40"
"\x80\x27\x88\xBA\x2C\x74\x42\xE0"
"\x1B\xA5\x85\x08\xB9\xE6\x22\x7A"
"\x36\x3B\x0D\x9F\xA0\x22\x6C\x2A"
"\x91\x75\x47\xBC\x67\x21\x4E\xF9"
"\xEA\xFF\xD9\xD5\xC0\xFC\x9E\x2C"
"\x3E\xAD\xC6\x61\x0E\x93\x7A\x22"
"\x09\xC8\x8D\xC1\x8E\xB4\x8B\x5C"
"\xC6\x24\x42\xB8\x23\x66\x80\xA9"
"\x32\x0B\x7A\x29\xBF\xB3\x0B\x63"
"\x43\x27\x13\xA9\xBE\xEB\xBD\xF3"
"\x33\x62\x70\xE2\x1B\x86\x7A\xA1"
"\x51\x4A\x16\xFE\x29\x63\x7E\xD0"
"\x7A\xA4\x6E\x2C\xF8\xC1\xDB\xE8"
"\xCB\x4D\xD2\x8C\x04\x14\xB4\x66"
"\x41\xB7\x3A\x96\x16\x7C\x1D\x5B"
"\xB6\x41\x42\x64\x43\xEE\x6E\x7C"
"\x8B\xAF\x01\x9C\xA4\x6E\x75\x8F"
"\xDE\x10\x9F\xA6\xE7\xD6\x44\x97"
"\x66\xA3\x96\x0F\x1C\x25\x60\xF5"
"\x3C\x2E\x32\x69\x0E\x82\xFF\x27"
"\x0F\xB5\x06\xDA\xD8\x31\x15\x6C"
"\xDF\x18\x6C\x87\xF5\x3B\x11\x9A"
"\x1B\x42\x1F\x5B\x29\x19\x96\x13"
"\x68\x2E\x5E\x08\x1C\x8F\x32\x4B"
"\x81\x77\x6D\xF4\xA0\x01\x42\xEC"
"\xDD\x5B\xFD\x3A\x8E\x6A\x14\xFB"
"\x83\x54\xDF\x0F\x86\xB7\xEA\x40"
"\x46\x39\xF7\x2A\x89\x8D\x4E\x96"
"\x5F\x5F\x6D\x76\xC6\x13\x9D\x3D"
"\x1D\x5F\x0C\x7D\xE2\xBC\xC2\x16"
"\x16\xBE\x89\x3E\xB0\x61\xA2\x5D"
"\xAF\xD1\x40\x5F\x1A\xB8\x26\x41"
"\xC6\xBD\x36\xEF\xED\x29\x50\x6D"
"\x10\xEF\x26\xE8\xA8\x93\x11\x3F"
"\x2D\x1F\x88\x20\x77\x45\xF5\x66"
"\x08\xB9\xF1\xEF\xB1\x93\xA8\x81"
"\x65\xC5\xCD\x3E\x8C\x06\x60\x2C"
"\xB2\x10\x7A\xCA\x05\x25\x59\xDB"
"\xC7\x28\xF5\x20\x35\x52\x9E\x62"
"\xF8\x88\x24\x1C\x4D\x84\x12\x39"
"\x39\xE4\x2E\xF4\xD4\x9D\x2B\xBC"
"\x87\x66\xE6\xC0\x6B\x31\x9A\x66"
"\x03\xDC\x95\xD8\x6B\xD0\x30\x8F"
"\xDF\x8F\x8D\xFA\xEC\x1F\x08\xBD"
"\xA3\x63\xE2\x71\x4F\x03\x94\x87"
"\x50\xDF\x15\x1F\xED\x3A\xA3\x7F"
"\x1F\x2A\xB5\xA1\x69\xAC\x4B\x0D"
"\x84\x9B\x2A\xE9\x55\xDD\x46\x91"
"\x15\x33\xF3\x2B\x9B\x46\x97\x00"
"\xF0\x29\xD8\x59\x5D\x33\x37\xF9"
"\x58\x33\x9B\x78\xC7\x58\x48\x6B"
"\x2C\x75\x64\xC4\xCA\xC1\x7E\xD5",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
};
static const struct cipher_testvec tf_cbc_tv_template[] = {
{ /* Generated with Nettle */
.key = zeroed_string,
.klen = 16,
.iv = zeroed_string,
.ptext = zeroed_string,
.ctext = "\x9f\x58\x9f\x5c\xf6\x12\x2c\x32"
"\xb6\xbf\xec\x2f\x2a\xe8\xc3\x5a",
.len = 16,
}, {
.key = zeroed_string,
.klen = 16,
.iv = "\x9f\x58\x9f\x5c\xf6\x12\x2c\x32"
"\xb6\xbf\xec\x2f\x2a\xe8\xc3\x5a",
.ptext = zeroed_string,
.ctext = "\xd4\x91\xdb\x16\xe7\xb1\xc3\x9e"
"\x86\xcb\x08\x6b\x78\x9f\x54\x19",
.len = 16,
}, {
.key = zeroed_string,
.klen = 16,
.iv = "\xd4\x91\xdb\x16\xe7\xb1\xc3\x9e"
"\x86\xcb\x08\x6b\x78\x9f\x54\x19",
.ptext = zeroed_string,
.ctext = "\x05\xef\x8c\x61\xa8\x11\x58\x26"
"\x34\xba\x5c\xb7\x10\x6a\xa6\x41",
.len = 16,
}, {
.key = zeroed_string,
.klen = 16,
.iv = zeroed_string,
.ptext = zeroed_string,
.ctext = "\x9f\x58\x9f\x5c\xf6\x12\x2c\x32"
"\xb6\xbf\xec\x2f\x2a\xe8\xc3\x5a"
"\xd4\x91\xdb\x16\xe7\xb1\xc3\x9e"
"\x86\xcb\x08\x6b\x78\x9f\x54\x19"
"\x05\xef\x8c\x61\xa8\x11\x58\x26"
"\x34\xba\x5c\xb7\x10\x6a\xa6\x41",
.len = 48,
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add blowfish test-vectors Add tests for parallel blowfish-x86_64 code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:03 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
crypto: testmgr - expand blowfish test vectors AVX2 implementation of blowfish cipher processes 32 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:52 -06:00
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\xC8\xFF\xF2\x53\xA6\x27\x09\xD1"
"\x33\x38\xC2\xC0\x0C\x14\x7E\xB5"
"\x26\x1B\x05\x0C\x05\x12\x3F\xC0"
"\xF9\x1C\x02\x28\x40\x96\x6F\xD0"
"\x3D\x32\xDF\xDA\x56\x00\x6E\xEE"
"\x5B\x2A\x72\x9D\xC2\x4D\x19\xBC"
"\x8C\x53\xFA\x87\x6F\xDD\x81\xA3"
"\xB1\xD3\x44\x65\xDF\xE7\x63\x38"
"\x4A\xFC\xDC\xEC\x3F\x26\x8E\xB8"
"\x43\xFC\xFE\x18\xB5\x11\x6D\x31"
"\x81\x8B\x0D\x75\xF6\x80\xEC\x84"
"\x04\xB9\xE6\x09\x63\xED\x39\xDB"
"\xC3\xF6\x14\xD6\x6E\x5E\x8B\xBD"
"\x3E\xFA\xD7\x98\x50\x6F\xD9\x63"
"\x02\xCD\x0D\x39\x4B\x0D\xEC\x80"
"\xE3\x6A\x17\xF4\xCC\xAD\xFF\x68"
"\x45\xDD\xC8\x83\x1D\x41\x96\x0D"
"\x91\x2E\x05\xD3\x59\x82\xE0\x43"
"\x90\x4F\xB9\xF7\xAD\x6B\x2E\xAF"
"\xA7\x84\x00\x53\xCD\x6F\xD1\x0C"
"\x4E\xF9\x5A\x23\xFB\xCA\xC7\xD3"
"\xA9\xAA\x9D\xB2\x3F\x66\xF1\xAC"
"\x25\x21\x8F\xF7\xEF\xF2\x6A\xDF"
"\xE8\xDA\x75\x1A\x8A\xF1\xDD\x38"
"\x1F\xF9\x3D\x68\x4A\xBB\x9E\x34"
"\x1F\x66\x1F\x9C\x2B\x54\xFF\x60"
"\x7F\x29\x4B\x55\x80\x8F\x4E\xA7"
"\xA6\x9A\x0A\xD9\x0D\x19\x00\xF8"
"\x1F\xBC\x0C\x40\x6B\xEC\x99\x25"
"\x94\x70\x74\x0E\x1D\xC5\xBC\x12"
"\xF3\x42\xBE\x95\xBF\xFB\x4E\x55"
"\x9A\xB9\xCE\x14\x16\x5B\xDC\xD3"
"\x75\x42\x62\x04\x31\x1F\x95\x7C"
"\x66\x1A\x97\xDC\x2F\x40\x5C\x39"
"\x78\xE6\x02\xDB\x49\xE1\xC6\x47"
"\xC2\x78\x9A\xBB\xF3\xBE\xCB\x93"
"\xD8\xB8\xE8\xBB\x8C\xB3\x9B\xA7"
"\xC2\x89\xF3\x91\x88\x83\x3D\xF0"
"\x29\xA2\xCD\xB5\x79\x16\xC2\x40"
"\x11\x03\x8E\x9C\xFD\xC9\x43\xC4"
"\xC2\x19\xF0\x4A\x32\xEF\x0C\x2B"
"\xD3\x2B\xE9\xD4\x4C\xDE\x95\xCF"
"\x04\x03\xD3\x2C\x7F\x82\xC8\xFA"
"\x0F\xD8\x7A\x39\x7B\x01\x41\x9C"
"\x78\xB6\xC9\xBF\xF9\x78\x57\x88"
"\xB1\xA5\xE1\xE0\xD9\x16\xD4\xC8"
"\xEE\xC4\xBE\x7B\x55\x59\x00\x48"
"\x1B\xBC\x14\xFA\x2A\x9D\xC9\x1C"
"\xFB\x28\x3F\x95\xDD\xB7\xD6\xCE"
"\x3A\x7F\x09\x0C\x0E\x69\x30\x7D"
"\xBC\x68\x9C\x91\x2A\x59\x57\x04"
"\xED\x1A\x1E\x00\xB1\x85\x92\x04"
"\x28\x8C\x0C\x3C\xC1\xD5\x12\xF7"
"\x4C\x3E\xB0\xE7\x86\x62\x68\x91"
"\xFC\xC4\xE2\xCE\xA6\xDC\x5E\x93"
"\x5D\x8D\x8C\x68\xB3\xB2\xB9\x64"
"\x16\xB8\xC8\x6F\xD8\xEE\x21\xBD"
"\xAC\x18\x0C\x7D\x0D\x05\xAB\xF1"
"\xFA\xDD\xE2\x48\xDF\x4C\x02\x39"
"\x69\xA1\x62\xBD\x49\x3A\x9D\x91"
"\x30\x70\x56\xA4\x37\xDD\x7C\xC0"
"\x0A\xA3\x30\x10\x26\x25\x41\x2C",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
};
static const struct cipher_testvec tf_ctr_tv_template[] = {
{ /* Generated with Crypto++ */
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\xDF\xDD\x69\xFA\xB0\x2E\xFD\xFE"
"\x70\x9E\xC5\x4B\xC9\xD4\xA1\x30"
"\x26\x9B\x89\xA1\xEE\x43\xE0\x52"
"\x55\x17\x4E\xC7\x0E\x33\x1F\xF1"
"\x9F\x8D\x40\x9F\x24\xFD\x92\xA0"
"\xBC\x8F\x35\xDD\x67\x38\xD8\xAA"
"\xCF\xF8\x48\xCA\xFB\xE4\x5C\x60"
"\x01\x41\x21\x12\x38\xAB\x52\x4F"
"\xA8\x57\x20\xE0\x21\x6A\x17\x0D"
"\x0E\xF9\x8E\x49\x42\x00\x3C\x94"
"\x14\xC0\xD0\x8D\x8A\x98\xEB\x29"
"\xEC\xAE\x96\x44\xC0\x3C\x48\xDC"
"\x29\x35\x25\x2F\xE7\x11\x6C\x68"
"\xC8\x67\x0A\x2F\xF4\x07\xBE\xF9"
"\x2C\x31\x87\x40\xAB\xB2\xB6\xFA"
"\xD2\xC9\x6D\x5C\x50\xE9\xE6\x7E"
"\xE3\x0A\xD2\xD5\x6D\x8D\x64\x9E"
"\x70\xCE\x03\x76\xDD\xE0\xF0\x8C"
"\x84\x86\x8B\x6A\xFE\xC7\xF9\x69"
"\x2E\xFE\xFC\xC2\xC4\x1A\x55\x58"
"\xB3\xBE\xE2\x7E\xED\x39\x42\x6C"
"\xB4\x42\x97\x9A\xEC\xE1\x0A\x06"
"\x02\xC5\x03\x9D\xC4\x48\x15\x66"
"\x35\x6A\xC2\xC9\xA2\x26\x30\xBB"
"\xDB\x2D\xC8\x08\x2B\xA0\x29\x1A"
"\x23\x61\x48\xEA\x80\x04\x27\xAA"
"\x69\x49\xE8\xE8\x4A\x83\x6B\x5A"
"\xCA\x7C\xD3\xB1\xB5\x0B\xCC\x23"
"\x74\x1F\xA9\x87\xCD\xED\xC0\x2D"
"\xBF\xEB\xCF\x16\x2D\x2A\x2E\x1D"
"\x96\xBA\x36\x11\x45\x41\xDA\xCE"
"\xA4\x48\x80\x8B\x06\xF4\x98\x89"
"\x8B\x23\x08\x53\xF4\xD4\x5A\x24"
"\x8B\xF8\x43\x73\xD1\xEE\xC4\xB0"
"\xF8\xFE\x09\x0C\x75\x05\x38\x0B"
"\x7C\x81\xDE\x9D\xE4\x61\x37\x63"
"\x63\xAD\x12\xD2\x04\xB9\xCE\x45"
"\x5A\x1A\x6E\xB3\x78\x2A\xA4\x74"
"\x86\xD0\xE3\xFF\xDA\x38\x9C\xB5"
"\xB8\xB1\xDB\x38\x2F\xC5\x6A\xB4"
"\xEB\x6E\x96\xE8\x43\x80\xB5\x51"
"\x61\x2D\x48\xAA\x07\x65\x11\x8C"
"\x48\xE3\x90\x7E\x78\x3A\xEC\x97"
"\x05\x3D\x84\xE7\x90\x2B\xAA\xBD"
"\x83\x29\x0E\x1A\x81\x73\x7B\xE0"
"\x7A\x01\x4A\x37\x3B\x77\x7F\x8D"
"\x49\xA4\x2F\x6E\xBE\x68\x99\x08"
"\x99\xAA\x4C\x12\x04\xAE\x1F\x77"
"\x35\x88\xF1\x65\x06\x0A\x0B\x4D"
"\x47\xF9\x50\x38\x5D\x71\xF9\x6E"
"\xDE\xEC\x61\x35\x2C\x4C\x96\x50"
"\xE8\x28\x93\x9C\x7E\x01\xC6\x04"
"\xB2\xD6\xBC\x6C\x17\xEB\xC1\x7D"
"\x11\xE9\x43\x83\x76\xAA\x53\x37"
"\x0C\x1D\x39\x89\x53\x72\x09\x7E"
"\xD9\x85\x16\x04\xA5\x2C\x05\x6F"
"\x17\x0C\x6E\x66\xAA\x84\xA7\xD9"
"\xE2\xD9\xC4\xEB\x43\x3E\xB1\x8D"
"\x7C\x36\xC7\x71\x70\x9C\x10\xD8"
"\xE8\x47\x2A\x4D\xFD\xA1\xBC\xE3"
"\xB9\x32\xE2\xC1\x82\xAC\xFE\xCC"
"\xC5\xC9\x7F\x9E\xCF\x33\x7A\xDF",
.len = 496,
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
}, { /* Generated with Crypto++ */
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
crypto: testmgr - expand twofish test vectors The AVX implementation of the twofish cipher processes 8 blocks parallel, so we need to make test vectors larger to check parallel code paths. Test vectors are also large enough to deal with 16 block parallel implementations which may occur in the future. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-05-28 07:55:38 -06:00
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xEB\x44\xAF\x49\x27\xB8\xFB\x44"
"\x4C\xA6\xC3\x0C\x8B\xD0\x01\x0C"
"\x53\xC8\x16\x38\xDE\x40\x4F\x91"
"\x25\x6D\x4C\xA0\x9A\x87\x1E\xDA"
"\x88\x7E\x89\xE9\x67\x2B\x83\xA2"
"\x5F\x2E\x23\x3E\x45\xB9\x77\x7B"
"\xA6\x7E\x47\x36\x81\x9F\x9B\xF3"
"\xE0\xF0\xD7\x47\xA9\xC8\xEF\x33"
"\x0C\x43\xFE\x67\x50\x0A\x2C\x3E"
"\xA0\xE1\x25\x8E\x80\x07\x4A\xC0"
"\x64\x89\x9F\x6A\x27\x96\x07\xA6"
"\x9B\xC8\x1B\x21\x60\xAE\x5D\x01"
"\xE2\xCD\xC8\xAA\x6C\x9D\x1C\x34"
"\x39\x18\x09\xA4\x82\x59\x78\xE7"
"\xFC\x59\x65\xF2\x94\xFF\xFB\xE2"
"\x3C\xDA\xB1\x90\x95\xBF\x91\xE3"
"\xE6\x87\x31\x9E\x16\x85\xAD\xB1"
"\x4C\xAE\x43\x4D\x19\x58\xB5\x5E"
"\x2E\xF5\x09\xAA\x39\xF4\xC0\xB3"
"\xD4\x4D\xDB\x73\x7A\xD4\xF1\xBF"
"\x89\x16\x4D\x2D\xA2\x26\x33\x72"
"\x18\x33\x7E\xD6\xD2\x16\xA4\x54"
"\xF4\x8C\xB3\x52\xDF\x21\x9C\xEB"
"\xBF\x49\xD3\xF9\x05\x06\xCB\xD2"
"\xA9\xD2\x3B\x6E\x19\x8C\xBC\x19"
"\xAB\x89\xD6\xD8\xCD\x56\x89\x5E"
"\xAC\x00\xE3\x50\x63\x4A\x80\x9A"
"\x05\xBC\x50\x39\xD3\x32\xD9\x0D"
"\xE3\x20\x0D\x75\x54\xEC\xE6\x31"
"\x14\xB9\x3A\x59\x00\x43\x37\x8E"
"\x8C\x5A\x79\x62\x14\x76\x8A\xAE"
"\x8F\xCC\xA1\x6C\x38\x78\xDD\x2D"
"\x8B\x6D\xEA\xBD\x7B\x25\xFF\x60"
"\xC9\x87\xB1\x79\x1E\xA5\x86\x68"
"\x81\xB4\xE2\xC1\x05\x7D\x3A\x73"
"\xD0\xDA\x75\x77\x9E\x05\x27\xF1"
"\x08\xA9\x66\x64\x6C\xBC\x82\x17"
"\x2C\x23\x5F\x62\x4D\x02\x1A\x58"
"\xE7\xB7\x23\x6D\xE2\x20\xDA\xEF"
"\xB4\xB3\x3F\xB2\x2B\x69\x98\x83"
"\x95\x87\x13\x57\x60\xD7\xB5\xB1"
"\xEE\x0A\x2F\x95\x36\x4C\x76\x5D"
"\x5F\xD9\x19\xED\xB9\xA5\x48\xBF"
"\xC8\xAB\x0F\x71\xCC\x61\x8E\x0A"
"\xD0\x29\x44\xA8\xB9\xC1\xE8\xC8"
"\xC9\xA8\x28\x81\xFB\x50\xF2\xF0"
"\x26\xAE\x39\xB8\x91\xCD\xA8\xAC"
"\xDE\x55\x1B\x50\x14\x53\x44\x17"
"\x54\x46\xFC\xB1\xE4\x07\x6B\x9A"
"\x01\x14\xF0\x2E\x2E\xDB\x46\x1B"
"\x1A\x09\x97\xA9\xB6\x97\x79\x06"
"\xFB\xCB\x85\xCF\xDD\xA1\x41\xB1"
"\x00\xAA\xF7\xE0\x89\x73\xFB\xE5"
"\xBF\x84\xDB\xC9\xCD\xC4\xA2\x0D"
"\x3B\xAC\xF9\xDF\x96\xBF\x88\x23"
"\x41\x67\xA1\x24\x99\x7E\xCC\x9B"
"\x02\x8F\x6A\x49\xF6\x25\xBA\x7A"
"\xF4\x78\xFD\x79\x62\x63\x4F\x14"
"\xD6\x11\x11\x04\x05\x5F\x7E\xEA"
"\x4C\xB6\xF8\xF4\x5F\x48\x52\x54"
"\x94\x63\xA8\x4E\xCF\xD2\x1B\x1B"
"\x22\x18\x6A\xAF\x6E\x3E\xE1\x0D",
.len = 496,
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
crypto: testmgr - expand twofish test vectors The AVX implementation of the twofish cipher processes 8 blocks parallel, so we need to make test vectors larger to check parallel code paths. Test vectors are also large enough to deal with 16 block parallel implementations which may occur in the future. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-05-28 07:55:38 -06:00
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59",
.ctext = "\xDF\xDD\x69\xFA\xB0\x2E\xFD\xFE"
"\x70\x9E\xC5\x4B\xC9\xD4\xA1\x30"
"\x26\x9B\x89\xA1\xEE\x43\xE0\x52"
"\x55\x17\x4E\xC7\x0E\x33\x1F\xF1"
"\x9F\x8D\x40\x9F\x24\xFD\x92\xA0"
"\xBC\x8F\x35\xDD\x67\x38\xD8\xAA"
"\xCF\xF8\x48\xCA\xFB\xE4\x5C\x60"
"\x01\x41\x21\x12\x38\xAB\x52\x4F"
"\xA8\x57\x20\xE0\x21\x6A\x17\x0D"
"\x0E\xF9\x8E\x49\x42\x00\x3C\x94"
"\x14\xC0\xD0\x8D\x8A\x98\xEB\x29"
"\xEC\xAE\x96\x44\xC0\x3C\x48\xDC"
"\x29\x35\x25\x2F\xE7\x11\x6C\x68"
"\xC8\x67\x0A\x2F\xF4\x07\xBE\xF9"
"\x2C\x31\x87\x40\xAB\xB2\xB6\xFA"
"\xD2\xC9\x6D\x5C\x50\xE9\xE6\x7E"
"\xE3\x0A\xD2\xD5\x6D\x8D\x64\x9E"
"\x70\xCE\x03\x76\xDD\xE0\xF0\x8C"
"\x84\x86\x8B\x6A\xFE\xC7\xF9\x69"
"\x2E\xFE\xFC\xC2\xC4\x1A\x55\x58"
"\xB3\xBE\xE2\x7E\xED\x39\x42\x6C"
"\xB4\x42\x97\x9A\xEC\xE1\x0A\x06"
"\x02\xC5\x03\x9D\xC4\x48\x15\x66"
"\x35\x6A\xC2\xC9\xA2\x26\x30\xBB"
"\xDB\x2D\xC8\x08\x2B\xA0\x29\x1A"
"\x23\x61\x48\xEA\x80\x04\x27\xAA"
"\x69\x49\xE8\xE8\x4A\x83\x6B\x5A"
"\xCA\x7C\xD3\xB1\xB5\x0B\xCC\x23"
"\x74\x1F\xA9\x87\xCD\xED\xC0\x2D"
"\xBF\xEB\xCF\x16\x2D\x2A\x2E\x1D"
"\x96\xBA\x36\x11\x45\x41\xDA\xCE"
"\xA4\x48\x80\x8B\x06\xF4\x98\x89"
"\x8B\x23\x08\x53\xF4\xD4\x5A\x24"
"\x8B\xF8\x43\x73\xD1\xEE\xC4\xB0"
"\xF8\xFE\x09\x0C\x75\x05\x38\x0B"
"\x7C\x81\xDE\x9D\xE4\x61\x37\x63"
"\x63\xAD\x12\xD2\x04\xB9\xCE\x45"
"\x5A\x1A\x6E\xB3\x78\x2A\xA4\x74"
"\x86\xD0\xE3\xFF\xDA\x38\x9C\xB5"
"\xB8\xB1\xDB\x38\x2F\xC5\x6A\xB4"
"\xEB\x6E\x96\xE8\x43\x80\xB5\x51"
"\x61\x2D\x48\xAA\x07\x65\x11\x8C"
"\x48\xE3\x90\x7E\x78\x3A\xEC\x97"
"\x05\x3D\x84\xE7\x90\x2B\xAA\xBD"
"\x83\x29\x0E\x1A\x81\x73\x7B\xE0"
"\x7A\x01\x4A\x37\x3B\x77\x7F\x8D"
"\x49\xA4\x2F\x6E\xBE\x68\x99\x08"
"\x99\xAA\x4C\x12\x04\xAE\x1F\x77"
"\x35\x88\xF1\x65\x06\x0A\x0B\x4D"
"\x47\xF9\x50\x38\x5D\x71\xF9\x6E"
"\xDE\xEC\x61\x35\x2C\x4C\x96\x50"
"\xE8\x28\x93\x9C\x7E\x01\xC6\x04"
"\xB2\xD6\xBC\x6C\x17\xEB\xC1\x7D"
"\x11\xE9\x43\x83\x76\xAA\x53\x37"
"\x0C\x1D\x39\x89\x53\x72\x09\x7E"
"\xD9\x85\x16\x04\xA5\x2C\x05\x6F"
"\x17\x0C\x6E\x66\xAA\x84\xA7\xD9"
"\xE2\xD9\xC4\xEB\x43\x3E\xB1\x8D"
"\x7C\x36\xC7\x71\x70\x9C\x10\xD8"
"\xE8\x47\x2A\x4D\xFD\xA1\xBC\xE3"
"\xB9\x32\xE2\xC1\x82\xAC\xFE\xCC"
"\xC5\xC9\x7F\x9E\xCF\x33\x7A\xDF"
"\x6C\x82\x9D",
.len = 499,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 2,
.tap = { 499 - 16, 16 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec tf_lrw_tv_template[] = {
/* Generated from AES-LRW test vectors */
{
.key = "\x45\x62\xac\x25\xf8\x28\x17\x6d"
"\x4c\x26\x84\x14\xb5\x68\x01\x85"
"\x25\x8e\x2a\x05\xe7\x3e\x9d\x03"
"\xee\x5a\x83\x0c\xcc\x09\x4c\x87",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xa1\x6c\x50\x69\x26\xa4\xef\x7b"
"\x7c\xc6\x91\xeb\x72\xdd\x9b\xee",
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x59\x70\x47\x14\xf5\x57\x47\x8c"
"\xd7\x79\xe8\x0f\x54\x88\x79\x44"
"\x0d\x48\xf0\xb7\xb1\x5a\x53\xea"
"\x1c\xaa\x6b\x29\xc2\xca\xfb\xaf",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xab\x72\x0a\xad\x3b\x0c\xf0\xc9"
"\x42\x2f\xf1\xae\xf1\x3c\xb1\xbd",
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd8\x2a\x91\x34\xb2\x6a\x56\x50"
"\x30\xfe\x69\xe2\x37\x7f\x98\x47"
"\xcd\xf9\x0b\x16\x0c\x64\x8f\xb6"
"\xb0\x0d\x0d\x1b\xae\x85\x87\x1f",
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x85\xa7\x56\x67\x08\xfa\x42\xe1"
"\x22\xe6\x82\xfc\xd9\xb4\xd7\xd4",
.len = 16,
}, {
.key = "\x0f\x6a\xef\xf8\xd3\xd2\xbb\x15"
"\x25\x83\xf7\x3c\x1f\x01\x28\x74"
"\xca\xc6\xbc\x35\x4d\x4a\x65\x54"
"\x90\xae\x61\xcf\x7b\xae\xbd\xcc"
"\xad\xe4\x94\xc5\x4a\x29\xae\x70",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xd2\xaf\x69\x35\x24\x1d\x0e\x1c"
"\x84\x8b\x05\xe4\xa2\x2f\x16\xf5",
.len = 16,
}, {
.key = "\x8a\xd4\xee\x10\x2f\xbd\x81\xff"
"\xf8\x86\xce\xac\x93\xc5\xad\xc6"
"\xa0\x19\x07\xc0\x9d\xf7\xbb\xdd"
"\x52\x13\xb2\xb7\xf0\xff\x11\xd8"
"\xd6\x08\xd0\xcd\x2e\xb1\x17\x6f",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x4a\x23\x56\xd7\xff\x90\xd0\x9a"
"\x0d\x7c\x26\xfc\xf0\xf0\xf6\xe4",
.len = 16,
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x30\xaf\x26\x05\x9d\x5d\x0a\x58"
"\xe2\xe7\xce\x8a\xb2\x56\x6d\x76",
.len = 16,
}, {
.key = "\xfb\x76\x15\xb2\x3d\x80\x89\x1d"
"\xd4\x70\x98\x0b\xc7\x95\x84\xc8"
"\xb2\xfb\x64\xce\x60\x97\x87\x8d"
"\x17\xfc\xe4\x5a\x49\xe8\x30\xb7"
"\x6e\x78\x17\xe7\x2d\x5e\x12\xd4"
"\x60\x64\x04\x7a\xf1\x2f\x9e\x0c",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xdf\xcf\xdc\xd2\xe1\xcf\x86\x75"
"\x17\x66\x5e\x0c\x14\xa1\x3d\x40",
.len = 16,
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x05\x11\xb7\x18\xab\xc6\x2d\xac"
"\x70\x5d\xf6\x22\x94\xcd\xe5\x6c"
"\x17\x6b\xf6\x1c\xf0\xf3\x6e\xf8"
"\x50\x38\x1f\x71\x49\xb6\x57\xd6"
"\x8f\xcb\x8d\x6b\xe3\xa6\x29\x90"
"\xfe\x2a\x62\x82\xae\x6d\x8b\xf6"
"\xad\x1e\x9e\x20\x5f\x38\xbe\x04"
"\xda\x10\x8e\xed\xa2\xa4\x87\xab"
"\xda\x6b\xb4\x0c\x75\xba\xd3\x7c"
"\xc9\xac\x42\x31\x95\x7c\xc9\x04"
"\xeb\xd5\x6e\x32\x69\x8a\xdb\xa6"
"\x15\xd7\x3f\x4f\x2f\x66\x69\x03"
"\x9c\x1f\x54\x0f\xde\x1f\xf3\x65"
"\x4c\x96\x12\xed\x7c\x92\x03\x01"
"\x6f\xbc\x35\x93\xac\xf1\x27\xf1"
"\xb4\x96\x82\x5a\x5f\xb0\xa0\x50"
"\x89\xa4\x8e\x66\x44\x85\xcc\xfd"
"\x33\x14\x70\xe3\x96\xb2\xc3\xd3"
"\xbb\x54\x5a\x1a\xf9\x74\xa2\xc5"
"\x2d\x64\x75\xdd\xb4\x54\xe6\x74"
"\x8c\xd3\x9d\x9e\x86\xab\x51\x53"
"\xb7\x93\x3e\x6f\xd0\x4e\x2c\x40"
"\xf6\xa8\x2e\x3e\x9d\xf4\x66\xa5"
"\x76\x12\x73\x44\x1a\x56\xd7\x72"
"\x88\xcd\x21\x8c\x4c\x0f\xfe\xda"
"\x95\xe0\x3a\xa6\xa5\x84\x46\xcd"
"\xd5\x3e\x9d\x3a\xe2\x67\xe6\x60"
"\x1a\xe2\x70\x85\x58\xc2\x1b\x09"
"\xe1\xd7\x2c\xca\xad\xa8\x8f\xf9"
"\xac\xb3\x0e\xdb\xca\x2e\xe2\xb8"
"\x51\x71\xd9\x3c\x6c\xf1\x56\xf8"
"\xea\x9c\xf1\xfb\x0c\xe6\xb7\x10"
"\x1c\xf8\xa9\x7c\xe8\x53\x35\xc1"
"\x90\x3e\x76\x4a\x74\xa4\x21\x2c"
"\xf6\x2c\x4e\x0f\x94\x3a\x88\x2e"
"\x41\x09\x6a\x33\x7d\xf6\xdd\x3f"
"\x8d\x23\x31\x74\x84\xeb\x88\x6e"
"\xcc\xb9\xbc\x22\x83\x19\x07\x22"
"\xa5\x2d\xdf\xa5\xf3\x80\x85\x78"
"\x84\x39\x6a\x6d\x6a\x99\x4f\xa5"
"\x15\xfe\x46\xb0\xe4\x6c\xa5\x41"
"\x3c\xce\x8f\x42\x60\x71\xa7\x75"
"\x08\x40\x65\x8a\x82\xbf\xf5\x43"
"\x71\x96\xa9\x4d\x44\x8a\x20\xbe"
"\xfa\x4d\xbb\xc0\x7d\x31\x96\x65"
"\xe7\x75\xe5\x3e\xfd\x92\x3b\xc9"
"\x55\xbb\x16\x7e\xf7\xc2\x8c\xa4"
"\x40\x1d\xe5\xef\x0e\xdf\xe4\x9a"
"\x62\x73\x65\xfd\x46\x63\x25\x3d"
"\x2b\xaf\xe5\x64\xfe\xa5\x5c\xcf"
"\x24\xf3\xb4\xac\x64\xba\xdf\x4b"
"\xc6\x96\x7d\x81\x2d\x8d\x97\xf7"
"\xc5\x68\x77\x84\x32\x2b\xcc\x85"
"\x74\x96\xf0\x12\x77\x61\xb9\xeb"
"\x71\xaa\x82\xcb\x1c\xdb\x89\xc8"
"\xc6\xb5\xe3\x5c\x7d\x39\x07\x24"
"\xda\x39\x87\x45\xc0\x2b\xbb\x01"
"\xac\xbc\x2a\x5c\x7f\xfc\xe8\xce"
"\x6d\x9c\x6f\xed\xd3\xc1\xa1\xd6"
"\xc5\x55\xa9\x66\x2f\xe1\xc8\x32"
"\xa6\x5d\xa4\x3a\x98\x73\xe8\x45"
"\xa4\xc7\xa8\xb4\xf6\x13\x03\xf6"
"\xe9\x2e\xc4\x29\x0f\x84\xdb\xc4"
"\x21\xc4\xc2\x75\x67\x89\x37\x0a",
.ctext = "\x30\x38\xeb\xaf\x12\x43\x1a\x89"
"\x62\xa2\x36\xe5\xcf\x77\x1e\xd9"
"\x08\xc3\x0d\xdd\x95\xab\x19\x96"
"\x27\x52\x41\xc3\xca\xfb\xf6\xee"
"\x40\x2d\xdf\xdd\x00\x0c\xb9\x0a"
"\x3a\xf0\xc0\xd1\xda\x63\x9e\x45"
"\x42\xe9\x29\xc0\xb4\x07\xb4\x31"
"\x66\x77\x72\xb5\xb6\xb3\x57\x46"
"\x34\x9a\xfe\x03\xaf\x6b\x36\x07"
"\x63\x8e\xc2\x5d\xa6\x0f\xb6\x7d"
"\xfb\x6d\x82\x51\xb6\x98\xd0\x71"
"\xe7\x10\x7a\xdf\xb2\xbd\xf1\x1d"
"\x72\x2b\x54\x13\xe3\x6d\x79\x37"
"\xa9\x39\x2c\xdf\x21\xab\x87\xd5"
"\xee\xef\x9a\x12\x50\x39\x2e\x1b"
"\x7d\xe6\x6a\x27\x48\xb9\xe7\xac"
"\xaa\xcd\x79\x5f\xf2\xf3\xa0\x08"
"\x6f\x2c\xf4\x0e\xd1\xb8\x89\x25"
"\x31\x9d\xef\xb1\x1d\x27\x55\x04"
"\xc9\x8c\xb7\x68\xdc\xb6\x67\x8a"
"\xdb\xcf\x22\xf2\x3b\x6f\xce\xbb"
"\x26\xbe\x4f\x27\x04\x42\xd1\x44"
"\x4c\x08\xa3\x95\x4c\x7f\x1a\xaf"
"\x1d\x28\x14\xfd\xb1\x1a\x34\x18"
"\xf5\x1e\x28\x69\x95\x6a\x5a\xba"
"\x8e\xb2\x58\x1d\x28\x17\x13\x3d"
"\x38\x7d\x14\x8d\xab\x5d\xf9\xe8"
"\x3c\x0f\x2b\x0d\x2b\x08\xb4\x4b"
"\x6b\x0d\xc8\xa7\x84\xc2\x3a\x1a"
"\xb7\xbd\xda\x92\x29\xb8\x5b\x5a"
"\x63\xa5\x99\x82\x09\x72\x8f\xc6"
"\xa4\x62\x24\x69\x8c\x2d\x26\x00"
"\x99\x83\x91\xd6\xc6\xcf\x57\x67"
"\x38\xea\xf2\xfc\x29\xe0\x73\x39"
"\xf9\x13\x94\x6d\xe2\x58\x28\x75"
"\x3e\xae\x71\x90\x07\x70\x1c\x38"
"\x5b\x4c\x1e\xb5\xa5\x3b\x20\xef"
"\xb1\x4c\x3e\x1a\x72\x62\xbb\x22"
"\x82\x09\xe3\x18\x3f\x4f\x48\xfc"
"\xdd\xac\xfc\xb6\x09\xdb\xd2\x7b"
"\xd6\xb7\x7e\x41\x2f\x14\xf5\x0e"
"\xc3\xac\x4a\xed\xe7\x82\xef\x31"
"\x1f\x1a\x51\x1e\x29\x60\xc8\x98"
"\x93\x51\x1d\x3d\x62\x59\x83\x82"
"\x0c\xf1\xd7\x8d\xac\x33\x44\x81"
"\x3c\x59\xb7\xd4\x5b\x65\x82\xc4"
"\xec\xdc\x24\xfd\x0e\x1a\x79\x94"
"\x34\xb0\x62\xfa\x98\x49\x26\x1f"
"\xf4\x9e\x40\x44\x5b\x1f\xf8\xbe"
"\x36\xff\xc6\xc6\x9d\xf2\xd6\xcc"
"\x63\x93\x29\xb9\x0b\x6d\xd7\x6c"
"\xdb\xf6\x21\x80\xf7\x5a\x37\x15"
"\x0c\xe3\x36\xc8\x74\x75\x20\x91"
"\xdf\x52\x2d\x0c\xe7\x45\xff\x46"
"\xb3\xf4\xec\xc2\xbd\xd3\x37\xb6"
"\x26\xa2\x5d\x7d\x61\xbf\x10\x46"
"\x57\x8d\x05\x96\x70\x0b\xd6\x41"
"\x5c\xe9\xd3\x54\x81\x39\x3a\xdd"
"\x5f\x92\x81\x6e\x35\x03\xd4\x72"
"\x3d\x5a\xe7\xb9\x3b\x0c\x84\x23"
"\x45\x5d\xec\x72\xc1\x52\xef\x2e"
"\x81\x00\xd3\xfe\x4c\x3c\x05\x61"
"\x80\x18\xc4\x6c\x03\xd3\xb7\xba"
"\x11\xd7\xb8\x6e\xea\xe1\x80\x30",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec tf_xts_tv_template[] = {
/* Generated from AES-XTS test vectors */
{
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x4b\xc9\x44\x4a\x11\xa3\xef\xac"
"\x30\x74\xe4\x44\x52\x77\x97\x43"
"\xa7\x60\xb2\x45\x2e\xf9\x00\x90"
"\x9f\xaa\xfd\x89\x6e\x9d\x4a\xe0",
.len = 32,
}, {
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\x57\x0e\x8f\xe5\x2a\x35\x61\x4f"
"\x32\xd3\xbd\x36\x05\x15\x44\x2c"
"\x58\x06\xf7\xf8\x00\xa8\xb6\xd5"
"\xc6\x28\x92\xdb\xd8\x34\xa2\xe9",
.len = 32,
}, {
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\x96\x45\x8f\x8d\x7a\x75\xb1\xde"
"\x40\x0c\x89\x56\xf6\x4d\xa7\x07"
"\x38\xbb\x5b\xe9\xcd\x84\xae\xb2"
"\x7b\x6a\x62\xf4\x8c\xb5\x37\xea",
.len = 32,
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ctext = "\xa9\x78\xae\x1e\xea\xa2\x44\x4c"
"\xa2\x7a\x64\x1f\xaf\x46\xc1\xe0"
"\x6c\xb2\xf3\x92\x9a\xd6\x7d\x58"
"\xb8\x2d\xb9\x5d\x58\x07\x66\x50"
"\xea\x35\x35\x8c\xb2\x46\x61\x06"
"\x5d\x65\xfc\x57\x8f\x69\x74\xab"
"\x8a\x06\x69\xb5\x6c\xda\x66\xc7"
"\x52\x90\xbb\x8e\x6d\x8b\xb5\xa2"
"\x78\x1d\xc2\xa9\xc2\x73\x00\xc3"
"\x32\x36\x7c\x97\x6b\x4e\x8a\x50"
"\xe4\x91\x83\x96\x8f\xf4\x94\x1a"
"\xa6\x27\xe1\x33\xcb\x91\xc6\x5f"
"\x94\x75\xbc\xd7\x3e\x3e\x6f\x9e"
"\xa9\x31\x80\x5e\xe5\xdb\xc8\x53"
"\x01\x73\x68\x32\x25\x19\xfa\xfb"
"\xe4\xcf\xb9\x3e\xa2\xa0\x8f\x31"
"\xbf\x54\x06\x93\xa8\xb1\x0f\xb6"
"\x7c\x3c\xde\x6f\x0f\xfb\x0c\x11"
"\x39\x80\x39\x09\x97\x65\xf2\x83"
"\xae\xe6\xa1\x6f\x47\xb8\x49\xde"
"\x99\x36\x20\x7d\x97\x3b\xec\xfa"
"\xb4\x33\x6e\x7a\xc7\x46\x84\x49"
"\x91\xcd\xe1\x57\x0d\xed\x40\x08"
"\x13\xf1\x4e\x3e\xa4\xa4\x5c\xe6"
"\xd2\x0c\x20\x8f\x3e\xdf\x3f\x47"
"\x9a\x2f\xde\x6d\x66\xc9\x99\x4a"
"\x2d\x9e\x9d\x4b\x1a\x27\xa2\x12"
"\x99\xf0\xf8\xb1\xb6\xf6\x57\xc3"
"\xca\x1c\xa3\x8e\xed\x39\x28\xb5"
"\x10\x1b\x4b\x08\x42\x00\x4a\xd3"
"\xad\x5a\xc6\x8e\xc8\xbb\x95\xc4"
"\x4b\xaa\xfe\xd5\x42\xa8\xa3\x6d"
"\x3c\xf3\x34\x91\x2d\xb4\xdd\x20"
"\x0c\x90\x6d\xa3\x9b\x66\x9d\x24"
"\x02\xa6\xa9\x3f\x3f\x58\x5d\x47"
"\x24\x65\x63\x7e\xbd\x8c\xe6\x52"
"\x7d\xef\x33\x53\x63\xec\xaa\x0b"
"\x64\x15\xa9\xa6\x1f\x10\x00\x38"
"\x35\xa8\xe7\xbe\x23\x70\x22\xe0"
"\xd3\xb9\xe6\xfd\xe6\xaa\x03\x50"
"\xf3\x3c\x27\x36\x8b\xcc\xfe\x9c"
"\x9c\xa3\xb3\xe7\x68\x9b\xa2\x71"
"\xe0\x07\xd9\x1f\x68\x1f\xac\x5e"
"\x7a\x74\x85\xa9\x6a\x90\xab\x2c"
"\x38\x51\xbc\x1f\x43\x4a\x56\x1c"
"\xf8\x47\x03\x4e\x67\xa8\x1f\x99"
"\x04\x39\x73\x32\xb2\x86\x79\xe7"
"\x14\x28\x70\xb8\xe2\x7d\x69\x85"
"\xb6\x0f\xc5\xd0\xd0\x01\x5c\xe6"
"\x09\x0f\x75\xf7\xb6\x81\xd2\x11"
"\x20\x9c\xa1\xee\x11\x44\x79\xd0"
"\xb2\x34\x77\xda\x10\x9a\x6f\x6f"
"\xef\x7c\xd9\xdc\x35\xb7\x61\xdd"
"\xf1\xa4\xc6\x1c\xbf\x05\x22\xac"
"\xfe\x2f\x85\x00\x44\xdf\x33\x16"
"\x35\xb6\xa3\xd3\x70\xdf\x69\x35"
"\x6a\xc7\xb4\x99\x45\x27\xc8\x8e"
"\x5a\x14\x30\xd0\x55\x3e\x4f\x64"
"\x0d\x38\xe3\xdf\x8b\xa8\x93\x26"
"\x75\xae\xf6\xb5\x23\x0b\x17\x31"
"\xbf\x27\xb8\xb5\x94\x31\xa7\x8f"
"\x43\xc4\x46\x24\x22\x4f\x8f\x7e"
"\xe5\xf4\x6d\x1e\x0e\x18\x7a\xbb"
"\xa6\x8f\xfb\x49\x49\xd8\x7e\x5a",
.len = 512,
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ctext = "\xd7\x4b\x93\x7d\x13\xa2\xa2\xe1"
"\x35\x39\x71\x88\x76\x1e\xc9\xea"
"\x86\xad\xf3\x14\x48\x3d\x5e\xe9"
"\xe9\x2d\xb2\x56\x59\x35\x9d\xec"
"\x84\xfa\x7e\x9d\x6d\x33\x36\x8f"
"\xce\xf4\xa9\x21\x0b\x5f\x96\xec"
"\xcb\xf9\x57\x68\x33\x88\x39\xbf"
"\x2f\xbb\x59\x03\xbd\x66\x8b\x11"
"\x11\x65\x51\x2e\xb8\x67\x05\xd1"
"\x27\x11\x5c\xd4\xcc\x97\xc2\xb3"
"\xa9\x55\xaf\x07\x56\xd1\xdc\xf5"
"\x85\xdc\x46\xe6\xf0\x24\xeb\x93"
"\x4d\xf0\x9b\xf5\x73\x1c\xda\x03"
"\x22\xc8\x3a\x4f\xb4\x19\x91\x09"
"\x54\x0b\xf6\xfe\x17\x3d\x1a\x53"
"\x72\x60\x79\xcb\x0e\x32\x8a\x77"
"\xd5\xed\xdb\x33\xd7\x62\x16\x69"
"\x63\xe0\xab\xb5\xf6\x9c\x5f\x3d"
"\x69\x35\x61\x86\xf8\x86\xb9\x89"
"\x6e\x59\x35\xac\xf6\x6b\x33\xa0"
"\xea\xef\x96\x62\xd8\xa9\xcf\x56"
"\xbf\xdb\x8a\xfd\xa1\x82\x77\x73"
"\x3d\x94\x4a\x49\x42\x6d\x08\x60"
"\xa1\xea\xab\xb6\x88\x13\x94\xb8"
"\x51\x98\xdb\x35\x85\xdf\xf6\xb9"
"\x8f\xcd\xdf\x80\xd3\x40\x2d\x72"
"\xb8\xb2\x6c\x02\x43\x35\x22\x2a"
"\x31\xed\xcd\x16\x19\xdf\x62\x0f"
"\x29\xcf\x87\x04\xec\x02\x4f\xe4"
"\xa2\xed\x73\xc6\x69\xd3\x7e\x89"
"\x0b\x76\x10\x7c\xd6\xf9\x6a\x25"
"\xed\xcc\x60\x5d\x61\x20\xc1\x97"
"\x56\x91\x57\x28\xbe\x71\x0d\xcd"
"\xde\xc4\x9e\x55\x91\xbe\xd1\x28"
"\x9b\x90\xeb\x73\xf3\x68\x51\xc6"
"\xdf\x82\xcc\xd8\x1f\xce\x5b\x27"
"\xc0\x60\x5e\x33\xd6\xa7\x20\xea"
"\xb2\x54\xc7\x5d\x6a\x3b\x67\x47"
"\xcf\xa0\xe3\xab\x86\xaf\xc1\x42"
"\xe6\xb0\x23\x4a\xaf\x53\xdf\xa0"
"\xad\x12\x32\x31\x03\xf7\x21\xbe"
"\x2d\xd5\x82\x42\xb6\x4a\x3d\xcd"
"\xd8\x81\x77\xa9\x49\x98\x6c\x09"
"\xc5\xa3\x61\x12\x62\x85\x6b\xcd"
"\xb3\xf4\x20\x0c\x41\xc4\x05\x37"
"\x46\x5f\xeb\x71\x8b\xf1\xaf\x6e"
"\xba\xf3\x50\x2e\xfe\xa8\x37\xeb"
"\xe8\x8c\x4f\xa4\x0c\xf1\x31\xc8"
"\x6e\x71\x4f\xa5\xd7\x97\x73\xe0"
"\x93\x4a\x2f\xda\x7b\xe0\x20\x54"
"\x1f\x8d\x85\x79\x0b\x7b\x5e\x75"
"\xb9\x07\x67\xcc\xc8\xe7\x21\x15"
"\xa7\xc8\x98\xff\x4b\x80\x1c\x12"
"\xa8\x54\xe1\x38\x52\xe6\x74\x81"
"\x97\x47\xa1\x41\x0e\xc0\x50\xe3"
"\x55\x0e\xc3\xa7\x70\x77\xce\x07"
"\xed\x8c\x88\xe6\xa1\x5b\x14\xec"
"\xe6\xde\x06\x6d\x74\xc5\xd9\xfa"
"\xe5\x2f\x5a\xff\xc8\x05\xee\x27"
"\x35\x61\xbf\x0b\x19\x78\x9b\xd2"
"\x04\xc7\x05\xb1\x79\xb4\xff\x5f"
"\xf3\xea\x67\x52\x78\xc2\xce\x70"
"\xa4\x05\x0b\xb2\xb3\xa8\x30\x97"
"\x37\x30\xe1\x91\x8d\xb3\x2a\xff",
.len = 512,
.also_non_np = 1,
.np = 3,
.tap = { 512 - 20, 4, 16 },
},
};
/*
* Serpent test vectors. These are backwards because Serpent writes
* octet sequences in right-to-left mode.
*/
static const struct cipher_testvec serpent_tv_template[] = {
{
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\x12\x07\xfc\xce\x9b\xd0\xd6\x47"
"\x6a\xe9\x8f\xbe\xd1\x43\xa0\xe2",
.len = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.klen = 16,
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\x4c\x7d\x8a\x32\x80\x72\xa2\x2c"
"\x82\x3e\x4a\x1f\x3a\xcd\xa1\x6d",
.len = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.klen = 32,
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\xde\x26\x9f\xf8\x33\xe4\x32\xb8"
"\x5b\x2e\x88\xd2\x70\x1c\xe7\x5c",
.len = 16,
}, {
.key = "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x80",
.klen = 16,
.ptext = zeroed_string,
.ctext = "\xdd\xd2\x6b\x98\xa5\xff\xd8\x2c"
"\x05\x34\x5a\x9d\xad\xbf\xaf\x49",
.len = 16,
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
crypto: testmgr - expand twofish test vectors The AVX implementation of the twofish cipher processes 8 blocks parallel, so we need to make test vectors larger to check parallel code paths. Test vectors are also large enough to deal with 16 block parallel implementations which may occur in the future. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-05-28 07:55:38 -06:00
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\xFB\xB0\x5D\xDE\xC0\xFE\xFC\xEB"
"\xB1\x80\x10\x43\xDE\x62\x70\xBD"
"\xFA\x8A\x93\xEA\x6B\xF7\xC5\xD7"
"\x0C\xD1\xBB\x29\x25\x14\x4C\x22"
"\x77\xA6\x38\x00\xDB\xB9\xE2\x07"
"\xD1\xAC\x82\xBA\xEA\x67\xAA\x39"
"\x99\x34\x89\x5B\x54\xE9\x12\x13"
"\x3B\x04\xE5\x12\x42\xC5\x79\xAB"
"\x0D\xC7\x3C\x58\x2D\xA3\x98\xF6"
"\xE4\x61\x9E\x17\x0B\xCE\xE8\xAA"
"\xB5\x6C\x1A\x3A\x67\x52\x81\x6A"
"\x04\xFF\x8A\x1B\x96\xFE\xE6\x87"
"\x3C\xD4\x39\x7D\x36\x9B\x03\xD5"
"\xB6\xA0\x75\x3C\x83\xE6\x1C\x73"
"\x9D\x74\x2B\x77\x53\x2D\xE5\xBD"
"\x69\xDA\x7A\x01\xF5\x6A\x70\x39"
"\x30\xD4\x2C\xF2\x8E\x06\x4B\x39"
"\xB3\x12\x1D\xB3\x17\x46\xE6\xD6"
"\xB6\x31\x36\x34\x38\x3C\x1D\x69"
"\x9F\x47\x28\x9A\x1D\x96\x70\x54"
"\x8E\x88\xCB\xE0\xF5\x6A\xAE\x0A"
"\x3C\xD5\x93\x1C\x21\xC9\x14\x3A"
"\x23\x9C\x9B\x79\xC7\x75\xC8\x39"
"\xA6\xAC\x65\x9A\x99\x37\xAF\x6D"
"\xBD\xB5\x32\xFD\xD8\x9C\x95\x7B"
"\xC6\x6A\x80\x64\xEA\xEF\x6D\x3F"
"\xA9\xFE\x5B\x16\xA3\xCF\x32\xC8"
"\xEF\x50\x22\x20\x93\x30\xBE\xE2"
"\x38\x05\x65\xAF\xBA\xB6\xE4\x72"
"\xA9\xEE\x05\x42\x88\xBD\x9D\x49"
"\xAD\x93\xCA\x4D\x45\x11\x43\x4D"
"\xB8\xF5\x74\x2B\x48\xE7\x21\xE4"
"\x4E\x3A\x4C\xDE\x65\x7A\x5A\xAD"
"\x86\xE6\x23\xEC\x6B\xA7\x17\xE6"
"\xF6\xA1\xAC\x29\xAE\xF9\x9B\x69"
"\x73\x65\x65\x51\xD6\x0B\x4E\x8C"
"\x17\x15\x9D\xB0\xCF\xB2\x42\x2B"
"\x51\xC3\x03\xE8\xB7\x7D\x2D\x39"
"\xE8\x10\x93\x16\xC8\x68\x4C\x60"
"\x87\x70\x14\xD0\x01\x57\xCB\x42"
"\x13\x59\xB1\x7F\x12\x4F\xBB\xC7"
"\xBD\x2B\xD4\xA9\x12\x26\x4F\xDE"
"\xFD\x72\xEC\xD7\x6F\x97\x14\x90"
"\x0E\x37\x13\xE6\x67\x1D\xE5\xFE"
"\x9E\x18\x3C\x8F\x3A\x3F\x59\x9B"
"\x71\x80\x05\x35\x3F\x40\x0B\x21"
"\x76\xE5\xEF\x42\x6C\xDB\x31\x05"
"\x5F\x05\xCF\x14\xE3\xF0\x61\xA2"
"\x49\x03\x5E\x77\x2E\x20\xBA\xA1"
"\xAF\x46\x51\xC0\x2B\xC4\x64\x1E"
"\x65\xCC\x51\x58\x0A\xDF\xF0\x5F"
"\x75\x9F\x48\xCD\x81\xEC\xC3\xF6"
"\xED\xC9\x4B\x7B\x4E\x26\x23\xE1"
"\xBB\xE9\x83\x0B\xCF\xE4\xDE\x00"
"\x48\xFF\xBF\x6C\xB4\x72\x16\xEF"
"\xC7\x46\xEE\x48\x8C\xB8\xAF\x45"
"\x91\x76\xE7\x6E\x65\x3D\x15\x86"
"\x10\xF8\xDB\x66\x97\x7C\x43\x4D"
"\x79\x12\x4E\xCE\x06\xD1\xD1\x6A"
"\x34\xC1\xC9\xF2\x28\x4A\xCD\x02"
"\x75\x55\x9B\xFF\x36\x73\xAB\x7C"
"\xF4\x46\x2E\xEB\xAC\xF3\xD2\xB7",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec tnepres_tv_template[] = {
{ /* KeySize=0 */
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\x41\xcc\x6b\x31\x59\x31\x45\x97"
"\x6d\x6f\xbb\x38\x4b\x37\x21\x28",
.len = 16,
},
{ /* KeySize=128, PT=0, I=1 */
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.key = "\x80\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.ctext = "\x49\xaf\xbf\xad\x9d\x5a\x34\x05"
"\x2c\xd8\xff\xa5\x98\x6b\xd2\xdd",
.len = 16,
}, { /* KeySize=128 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.klen = 16,
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\xea\xf4\xd7\xfc\xd8\x01\x34\x47"
"\x81\x45\x0b\xfa\x0c\xd6\xad\x6e",
.len = 16,
}, { /* KeySize=128, I=121 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x80",
.klen = 16,
.ptext = zeroed_string,
.ctext = "\x3d\xda\xbf\xc0\x06\xda\xab\x06"
"\x46\x2a\xf4\xef\x81\x54\x4e\x26",
.len = 16,
}, { /* KeySize=192, PT=0, I=1 */
.key = "\x80\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 24,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\xe7\x8e\x54\x02\xc7\x19\x55\x68"
"\xac\x36\x78\xf7\xa3\xf6\x0c\x66",
.len = 16,
}, { /* KeySize=256, PT=0, I=1 */
.key = "\x80\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\xab\xed\x96\xe7\x66\xbf\x28\xcb"
"\xc0\xeb\xd2\x1a\x82\xef\x08\x19",
.len = 16,
}, { /* KeySize=256, I=257 */
.key = "\x1f\x1e\x1d\x1c\x1b\x1a\x19\x18"
"\x17\x16\x15\x14\x13\x12\x11\x10"
"\x0f\x0e\x0d\x0c\x0b\x0a\x09\x08"
"\x07\x06\x05\x04\x03\x02\x01\x00",
.klen = 32,
.ptext = "\x0f\x0e\x0d\x0c\x0b\x0a\x09\x08"
"\x07\x06\x05\x04\x03\x02\x01\x00",
.ctext = "\x5c\xe7\x1c\x70\xd2\x88\x2e\x5b"
"\xb8\x32\xe4\x33\xf8\x9f\x26\xde",
.len = 16,
}, { /* KeySize=256 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.klen = 32,
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ctext = "\x64\xa9\x1a\x37\xed\x9f\xe7\x49"
"\xa8\x4e\x76\xd6\xf5\x0d\x78\xee",
.len = 16,
}
};
static const struct cipher_testvec serpent_cbc_tv_template[] = {
{ /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
crypto: testmgr - expand twofish test vectors The AVX implementation of the twofish cipher processes 8 blocks parallel, so we need to make test vectors larger to check parallel code paths. Test vectors are also large enough to deal with 16 block parallel implementations which may occur in the future. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-05-28 07:55:38 -06:00
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x80\xCF\x11\x41\x1A\xB9\x4B\x9C"
"\xFF\xB7\x6C\xEA\xF0\xAF\x77\x6E"
"\x71\x75\x95\x9D\x4E\x1C\xCF\xAD"
"\x81\x34\xE9\x8F\xAE\x5A\x91\x1C"
"\x38\x63\x35\x7E\x79\x18\x0A\xE8"
"\x67\x06\x76\xD5\xFF\x22\x2F\xDA"
"\xB6\x2D\x57\x13\xB6\x3C\xBC\x97"
"\xFE\x53\x75\x35\x97\x7F\x51\xEA"
"\xDF\x5D\xE8\x9D\xCC\xD9\xAE\xE7"
"\x62\x67\xFF\x04\xC2\x18\x22\x5F"
"\x2E\x06\xC1\xE2\x26\xCD\xC6\x1E"
"\xE5\x2C\x4E\x87\x23\xDD\xF0\x41"
"\x08\xA5\xB4\x3E\x07\x1E\x0B\xBB"
"\x72\x84\xF8\x0A\x3F\x38\x5E\x91"
"\x15\x26\xE1\xDB\xA4\x3D\x74\xD2"
"\x41\x1E\x3F\xA9\xC6\x7D\x2A\xAB"
"\x27\xDF\x89\x1D\x86\x3E\xF7\x5A"
"\xF6\xE3\x0F\xC7\x6B\x4C\x96\x7C"
"\x2D\x12\xA5\x05\x92\xCB\xD7\x4A"
"\x4D\x1E\x88\x21\xE1\x63\xB4\xFC"
"\x4A\xF2\xCD\x35\xB9\xD7\x70\x97"
"\x5A\x5E\x7E\x96\x52\x20\xDC\x25"
"\xE9\x6B\x36\xB4\xE0\x98\x85\x2C"
"\x3C\xD2\xF7\x78\x8A\x73\x26\x9B"
"\xAF\x0B\x11\xE8\x4D\x67\x23\xE9"
"\x77\xDF\x58\xF6\x6F\x9E\xA4\xC5"
"\x10\xA1\x82\x0E\x80\xA0\x8F\x4B"
"\xA1\xC0\x12\x54\x4E\xC9\x20\x92"
"\x11\x00\x10\x4E\xB3\x7C\xCA\x63"
"\xE5\x3F\xD3\x41\x37\xCD\x74\xB7"
"\xA5\x7C\x61\xB8\x0B\x7A\x7F\x4D"
"\xFE\x96\x7D\x1B\xBE\x60\x37\xB7"
"\x81\x92\x66\x67\x15\x1E\x39\x98"
"\x52\xC0\xF4\x69\xC0\x99\x4F\x5A"
"\x2E\x32\xAD\x7C\x8B\xE9\xAD\x05"
"\x55\xF9\x0A\x1F\x97\x5C\xFA\x2B"
"\xF4\x99\x76\x3A\x6E\x4D\xE1\x4C"
"\x14\x4E\x6F\x87\xEE\x1A\x85\xA3"
"\x96\xC6\x66\x49\xDA\x0D\x71\xAC"
"\x04\x05\x46\xD3\x90\x0F\x64\x64"
"\x01\x66\x2C\x62\x5D\x34\xD1\xCB"
"\x3A\x24\xCE\x95\xEF\xAE\x2C\x97"
"\x0E\x0C\x1D\x36\x49\xEB\xE9\x3D"
"\x62\xA6\x19\x28\x9E\x26\xB4\x3F"
"\xD7\x55\x42\x3C\xCD\x72\x0A\xF0"
"\x7D\xE9\x95\x45\x86\xED\xB1\xE0"
"\x8D\xE9\xC5\x86\x13\x24\x28\x7D"
"\x74\xEF\xCA\x50\x12\x7E\x64\x8F"
"\x1B\xF5\x5B\xFE\xE2\xAC\xFA\xE7"
"\xBD\x38\x8C\x11\x20\xEF\xB1\xAA"
"\x7B\xE5\xE5\x78\xAD\x9D\x2D\xA2"
"\x8E\xDD\x48\xB3\xEF\x18\x92\x7E"
"\xE6\x75\x0D\x54\x64\x11\xA3\x3A"
"\xDB\x97\x0F\xD3\xDF\x07\xD3\x7E"
"\x1E\xD1\x87\xE4\x74\xBB\x46\xF4"
"\xBA\x23\x2D\x8D\x29\x07\x12\xCF"
"\x34\xCD\x72\x7F\x01\x30\xE7\xA0"
"\xF8\xDD\xA8\x08\xF0\xBC\xB1\xA2"
"\xCC\xE1\x6B\x5F\xBE\xEA\xF1\xE4"
"\x02\xC4\xAF\xFA\xAD\x31\xF4\xBF"
"\xFC\x66\xAA\x37\xF2\x37\x39\x6B"
"\xBC\x08\x3A\xA2\x29\xB3\xDF\xD1",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
};
static const struct cipher_testvec serpent_ctr_tv_template[] = {
{ /* Generated with Crypto++ */
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x84\x68\xEC\xF2\x1C\x88\x20\xCA"
"\x37\x69\xE3\x3A\x22\x85\x48\x46"
"\x70\xAA\x25\xB4\xCD\x8B\x04\x4E"
"\x8D\x15\x2B\x98\xDF\x7B\x6D\xB9"
"\xE0\x4A\x73\x00\x65\xB6\x1A\x0D"
"\x5C\x60\xDF\x34\xDC\x60\x4C\xDF"
"\xB5\x1F\x26\x8C\xDA\xC1\x11\xA8"
"\x80\xFA\x37\x7A\x89\xAA\xAE\x7B"
"\x92\x6E\xB9\xDC\xC9\x62\x4F\x88"
"\x0A\x5D\x97\x2F\x6B\xAC\x03\x7C"
"\x22\xF6\x55\x5A\xFA\x35\xA5\x17"
"\xA1\x5C\x5E\x2B\x63\x2D\xB9\x91"
"\x3E\x83\x26\x00\x4E\xD5\xBE\xCE"
"\x79\xC4\x3D\xFC\x70\xA0\xAD\x96"
"\xBA\x58\x2A\x1C\xDF\xC2\x3A\xA5"
"\x7C\xB5\x12\x89\xED\xBF\xB6\x09"
"\x13\x4F\x7D\x61\x3C\x5C\x27\xFC"
"\x5D\xE1\x4F\xA1\xEA\xB3\xCA\xB9"
"\xE6\xD0\x97\x81\xDE\xD1\xFB\x8A"
"\x30\xDB\xA3\x5D\xEC\x25\x0B\x86"
"\x71\xC8\xA7\x67\xE8\xBC\x7D\x4C"
"\xAE\x82\xD3\x73\x31\x09\xCB\xB3"
"\x4D\xD4\xC0\x8A\x2B\xFA\xA6\x55"
"\x39\x0A\xBC\x6E\x75\xAB\xC2\xE2"
"\x8A\xF2\x26\xCD\x63\x38\x35\xF7"
"\xAE\x12\x83\xCD\x8A\x9E\x7E\x4C"
"\xFE\x4D\xD7\xCE\x5C\x6E\x4C\xAF"
"\xE3\xCD\x76\xA7\x87\xA1\x54\x7C"
"\xEC\x32\xC7\x83\x2A\xFF\xF8\xEA"
"\x87\xB2\x47\xA3\x9D\xC2\x9C\xA2"
"\xB7\x2C\x7C\x1A\x24\xCB\x88\x61"
"\xFF\xA7\x1A\x16\x01\xDD\x4B\xFC"
"\x2E\xE0\x48\x67\x09\x42\xCC\x91"
"\xBE\x20\x38\xC0\x5E\x3B\x95\x00"
"\xA1\x96\x66\x0B\x8A\xE9\x9E\xF7"
"\x6B\x34\x0A\x51\xC0\x3B\xEB\x71"
"\x07\x97\x38\x4B\x5C\x56\x98\x67"
"\x78\x9C\xD0\x0E\x2B\xB5\x67\x90"
"\x75\xF8\xFE\x6D\x4E\x85\xCC\x0D"
"\x18\x06\x15\x9D\x5A\x10\x13\x37"
"\xA3\xD6\x68\xA2\xDF\x7E\xC7\x12"
"\xC9\x0D\x4D\x91\xB0\x2A\x55\xFF"
"\x6F\x73\x13\xDF\x28\xB5\x2A\x2C"
"\xE4\xFC\x20\xD9\xF1\x7A\x82\xB1"
"\xCB\x57\xB6\x3D\x8C\xF4\x8E\x27"
"\x37\xDC\x35\xF3\x79\x01\x53\xA4"
"\x7B\x37\xDE\x7C\x04\xAE\x50\xDB"
"\x9B\x1E\x8C\x07\xA7\x52\x49\x50"
"\x34\x25\x65\xDD\xA9\x8F\x7E\xBD"
"\x7A\xC9\x36\xAE\xDE\x21\x48\x64"
"\xC2\x02\xBA\xBE\x11\x1E\x3D\x9C"
"\x98\x52\xCC\x04\xBD\x5E\x61\x26"
"\x10\xD3\x21\xD9\x6E\x25\x98\x77"
"\x8E\x98\x63\xF6\xF6\x52\xFB\x13"
"\xAA\x30\xF2\xB9\xA4\x43\x53\x39"
"\x1C\x97\x07\x7E\x6B\xFF\x3D\x43"
"\xA6\x71\x6B\x66\x8F\x58\x3F\x71"
"\x90\x47\x40\x92\xE6\x69\xD1\x96"
"\x34\xB3\x3B\xE5\x43\xE4\xD5\x56"
"\xB2\xE6\x7E\x86\x7A\x12\x17\x5B"
"\x30\xF3\x9B\x0D\xFA\x57\xE4\x50"
"\x40\x53\x77\x8C\x15\xF8\x8D\x13",
.len = 496,
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add twofish tests Add tests for parallel twofish-x86_64-3way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-10 14:03:12 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
crypto: testmgr - expand twofish test vectors The AVX implementation of the twofish cipher processes 8 blocks parallel, so we need to make test vectors larger to check parallel code paths. Test vectors are also large enough to deal with 16 block parallel implementations which may occur in the future. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-05-28 07:55:38 -06:00
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x84\x68\xEC\xF2\x1C\x88\x20\xCA"
"\x37\x69\xE3\x3A\x22\x85\x48\x46"
"\x70\xAA\x25\xB4\xCD\x8B\x04\x4E"
"\x8D\x15\x2B\x98\xDF\x7B\x6D\xB9"
"\xE0\x4A\x73\x00\x65\xB6\x1A\x0D"
"\x5C\x60\xDF\x34\xDC\x60\x4C\xDF"
"\xB5\x1F\x26\x8C\xDA\xC1\x11\xA8"
"\x80\xFA\x37\x7A\x89\xAA\xAE\x7B"
"\x92\x6E\xB9\xDC\xC9\x62\x4F\x88"
"\x0A\x5D\x97\x2F\x6B\xAC\x03\x7C"
"\x22\xF6\x55\x5A\xFA\x35\xA5\x17"
"\xA1\x5C\x5E\x2B\x63\x2D\xB9\x91"
"\x3E\x83\x26\x00\x4E\xD5\xBE\xCE"
"\x79\xC4\x3D\xFC\x70\xA0\xAD\x96"
"\xBA\x58\x2A\x1C\xDF\xC2\x3A\xA5"
"\x7C\xB5\x12\x89\xED\xBF\xB6\x09"
"\x13\x4F\x7D\x61\x3C\x5C\x27\xFC"
"\x5D\xE1\x4F\xA1\xEA\xB3\xCA\xB9"
"\xE6\xD0\x97\x81\xDE\xD1\xFB\x8A"
"\x30\xDB\xA3\x5D\xEC\x25\x0B\x86"
"\x71\xC8\xA7\x67\xE8\xBC\x7D\x4C"
"\xAE\x82\xD3\x73\x31\x09\xCB\xB3"
"\x4D\xD4\xC0\x8A\x2B\xFA\xA6\x55"
"\x39\x0A\xBC\x6E\x75\xAB\xC2\xE2"
"\x8A\xF2\x26\xCD\x63\x38\x35\xF7"
"\xAE\x12\x83\xCD\x8A\x9E\x7E\x4C"
"\xFE\x4D\xD7\xCE\x5C\x6E\x4C\xAF"
"\xE3\xCD\x76\xA7\x87\xA1\x54\x7C"
"\xEC\x32\xC7\x83\x2A\xFF\xF8\xEA"
"\x87\xB2\x47\xA3\x9D\xC2\x9C\xA2"
"\xB7\x2C\x7C\x1A\x24\xCB\x88\x61"
"\xFF\xA7\x1A\x16\x01\xDD\x4B\xFC"
"\x2E\xE0\x48\x67\x09\x42\xCC\x91"
"\xBE\x20\x38\xC0\x5E\x3B\x95\x00"
"\xA1\x96\x66\x0B\x8A\xE9\x9E\xF7"
"\x6B\x34\x0A\x51\xC0\x3B\xEB\x71"
"\x07\x97\x38\x4B\x5C\x56\x98\x67"
"\x78\x9C\xD0\x0E\x2B\xB5\x67\x90"
"\x75\xF8\xFE\x6D\x4E\x85\xCC\x0D"
"\x18\x06\x15\x9D\x5A\x10\x13\x37"
"\xA3\xD6\x68\xA2\xDF\x7E\xC7\x12"
"\xC9\x0D\x4D\x91\xB0\x2A\x55\xFF"
"\x6F\x73\x13\xDF\x28\xB5\x2A\x2C"
"\xE4\xFC\x20\xD9\xF1\x7A\x82\xB1"
"\xCB\x57\xB6\x3D\x8C\xF4\x8E\x27"
"\x37\xDC\x35\xF3\x79\x01\x53\xA4"
"\x7B\x37\xDE\x7C\x04\xAE\x50\xDB"
"\x9B\x1E\x8C\x07\xA7\x52\x49\x50"
"\x34\x25\x65\xDD\xA9\x8F\x7E\xBD"
"\x7A\xC9\x36\xAE\xDE\x21\x48\x64"
"\xC2\x02\xBA\xBE\x11\x1E\x3D\x9C"
"\x98\x52\xCC\x04\xBD\x5E\x61\x26"
"\x10\xD3\x21\xD9\x6E\x25\x98\x77"
"\x8E\x98\x63\xF6\xF6\x52\xFB\x13"
"\xAA\x30\xF2\xB9\xA4\x43\x53\x39"
"\x1C\x97\x07\x7E\x6B\xFF\x3D\x43"
"\xA6\x71\x6B\x66\x8F\x58\x3F\x71"
"\x90\x47\x40\x92\xE6\x69\xD1\x96"
"\x34\xB3\x3B\xE5\x43\xE4\xD5\x56"
"\xB2\xE6\x7E\x86\x7A\x12\x17\x5B"
"\x30\xF3\x9B\x0D\xFA\x57\xE4\x50"
"\x40\x53\x77\x8C\x15\xF8\x8D\x13"
"\x38\xE2\xE5",
.len = 499,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
.np = 2,
.tap = { 499 - 16, 16 },
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\x06\x9A\xF8\xB4\x53\x88\x62\xFC"
"\x68\xB8\x2E\xDF\xC1\x05\x0F\x3D"
"\xAF\x4D\x95\xAE\xC4\xE9\x1C\xDC"
"\xF6\x2B\x8F\x90\x89\xF6\x7E\x1A"
"\xA6\xB9\xE4\xF4\xFA\xCA\xE5\x7E"
"\x71\x28\x06\x4F\xE8\x08\x39\xDA"
"\xA5\x0E\xC8\xC0\xB8\x16\xE5\x69"
"\xE5\xCA\xEC\x4F\x63\x2C\xC0\x9B"
"\x9F\x3E\x39\x79\xF0\xCD\x64\x35"
"\x4A\xD3\xC8\xA9\x31\xCD\x48\x5B"
"\x92\x3D\x8F\x3F\x96\xBD\xB3\x18"
"\x74\x2A\x5D\x29\x3F\x57\x8F\xE2"
"\x67\x9A\xE0\xE5\xD4\x4A\xE2\x47"
"\xBC\xF6\xEB\x14\xF3\x8C\x20\xC2"
"\x7D\xE2\x43\x81\x86\x72\x2E\xB1"
"\x39\xF6\x95\xE1\x1F\xCB\x76\x33"
"\x5B\x7D\x23\x0F\x3A\x67\x2A\x2F"
"\xB9\x37\x9D\xDD\x1F\x16\xA1\x3C"
"\x70\xFE\x52\xAA\x93\x3C\xC4\x46"
"\xB1\xE5\xFF\xDA\xAF\xE2\x84\xFE"
"\x25\x92\xB2\x63\xBD\x49\x77\xB4"
"\x22\xA4\x6A\xD5\x04\xE0\x45\x58"
"\x1C\x34\x96\x7C\x03\x0C\x13\xA2"
"\x05\x22\xE2\xCB\x5A\x35\x03\x09"
"\x40\xD2\x82\x05\xCA\x58\x73\xF2"
"\x29\x5E\x01\x47\x13\x32\x78\xBE"
"\x06\xB0\x51\xDB\x6C\x31\xA0\x1C"
"\x74\xBC\x8D\x25\xDF\xF8\x65\xD1"
"\x38\x35\x11\x26\x4A\xB4\x06\x32"
"\xFA\xD2\x07\x77\xB3\x74\x98\x80"
"\x61\x59\xA8\x9F\xF3\x6F\x2A\xBF"
"\xE6\xA5\x9A\xC4\x6B\xA6\x49\x6F"
"\xBC\x47\xD9\xFB\xC6\xEF\x25\x65"
"\x96\xAC\x9F\xE4\x81\x4B\xD8\xBA"
"\xD6\x9B\xC9\x6D\x58\x40\x81\x02"
"\x73\x44\x4E\x43\x6E\x37\xBB\x11"
"\xE3\xF9\xB8\x2F\xEC\x76\x34\xEA"
"\x90\xCD\xB7\x2E\x0E\x32\x71\xE8"
"\xBB\x4E\x0B\x98\xA4\x17\x17\x5B"
"\x07\xB5\x82\x3A\xC4\xE8\x42\x51"
"\x5A\x4C\x4E\x7D\xBF\xC4\xC0\x4F"
"\x68\xB8\xC6\x4A\x32\x6F\x0B\xD7"
"\x85\xED\x6B\xFB\x72\xD2\xA5\x8F"
"\xBF\xF9\xAC\x59\x50\xA8\x08\x70"
"\xEC\xBD\x0A\xBF\xE5\x87\xA1\xC2"
"\x92\x14\x78\xAF\xE8\xEA\x2E\xDD"
"\xC1\x03\x9A\xAA\x89\x8B\x32\x46"
"\x5B\x18\x27\xBA\x46\xAA\x64\xDE"
"\xE3\xD5\xA3\xFC\x7B\x5B\x61\xDB"
"\x7E\xDA\xEC\x30\x17\x19\xF8\x80"
"\xB5\x5E\x27\xB5\x37\x3A\x1F\x28"
"\x07\x73\xC3\x63\xCE\xFF\x8C\xFE"
"\x81\x4E\xF8\x24\xF3\xB8\xC7\xE8"
"\x16\x9A\xCC\x58\x2F\x88\x1C\x4B"
"\xBB\x33\xA2\x73\xF0\x1C\x89\x0E"
"\xDC\x34\x27\x89\x98\xCE\x1C\xA2"
"\xD8\xB8\x90\xBE\xEC\x72\x28\x13"
"\xAC\x7B\xF1\xD0\x7F\x7A\x28\x50"
"\xB7\x99\x65\x8A\xC9\xC6\x21\x34"
"\x7F\x67\x9D\xB7\x2C\xCC\xF5\x17"
"\x2B\x89\xAC\xB0\xD7\x1E\x47\xB0"
"\x61\xAF\xD4\x63\x6D\xB8\x2D\x20",
.len = 496,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec serpent_lrw_tv_template[] = {
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
/* Generated from AES-LRW test vectors */
{
.key = "\x45\x62\xac\x25\xf8\x28\x17\x6d"
"\x4c\x26\x84\x14\xb5\x68\x01\x85"
"\x25\x8e\x2a\x05\xe7\x3e\x9d\x03"
"\xee\x5a\x83\x0c\xcc\x09\x4c\x87",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x6f\xbf\xd4\xa4\x5d\x71\x16\x79"
"\x63\x9c\xa6\x8e\x40\xbe\x0d\x8a",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\x59\x70\x47\x14\xf5\x57\x47\x8c"
"\xd7\x79\xe8\x0f\x54\x88\x79\x44"
"\x0d\x48\xf0\xb7\xb1\x5a\x53\xea"
"\x1c\xaa\x6b\x29\xc2\xca\xfb\xaf",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xfd\xb2\x66\x98\x80\x96\x55\xad"
"\x08\x94\x54\x9c\x21\x7c\x69\xe3",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\xd8\x2a\x91\x34\xb2\x6a\x56\x50"
"\x30\xfe\x69\xe2\x37\x7f\x98\x47"
"\xcd\xf9\x0b\x16\x0c\x64\x8f\xb6"
"\xb0\x0d\x0d\x1b\xae\x85\x87\x1f",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x14\x5e\x3d\x70\xc0\x6e\x9c\x34"
"\x5b\x5e\xcf\x0f\xe4\x8c\x21\x5c",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\x0f\x6a\xef\xf8\xd3\xd2\xbb\x15"
"\x25\x83\xf7\x3c\x1f\x01\x28\x74"
"\xca\xc6\xbc\x35\x4d\x4a\x65\x54"
"\x90\xae\x61\xcf\x7b\xae\xbd\xcc"
"\xad\xe4\x94\xc5\x4a\x29\xae\x70",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x25\x39\xaa\xa5\xf0\x65\xc8\xdc"
"\x5d\x45\x95\x30\x8f\xff\x2f\x1b",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\x8a\xd4\xee\x10\x2f\xbd\x81\xff"
"\xf8\x86\xce\xac\x93\xc5\xad\xc6"
"\xa0\x19\x07\xc0\x9d\xf7\xbb\xdd"
"\x52\x13\xb2\xb7\xf0\xff\x11\xd8"
"\xd6\x08\xd0\xcd\x2e\xb1\x17\x6f",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x0c\x20\x20\x63\xd6\x8b\xfc\x8f"
"\xc0\xe2\x17\xbb\xd2\x59\x6f\x26",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc1\x35\x2e\x53\xf0\x96\x4d\x9c"
"\x2e\x18\xe6\x99\xcd\xd3\x15\x68",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\xfb\x76\x15\xb2\x3d\x80\x89\x1d"
"\xd4\x70\x98\x0b\xc7\x95\x84\xc8"
"\xb2\xfb\x64\xce\x60\x97\x87\x8d"
"\x17\xfc\xe4\x5a\x49\xe8\x30\xb7"
"\x6e\x78\x17\xe7\x2d\x5e\x12\xd4"
"\x60\x64\x04\x7a\xf1\x2f\x9e\x0c",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x86\x0a\xc6\xa9\x1a\x9f\xe7\xe6"
"\x64\x3b\x33\xd6\xd5\x84\xd6\xdf",
.len = 16,
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\x11\xb7\x18\xab\xc6\x2d\xac"
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
"\x70\x5d\xf6\x22\x94\xcd\xe5\x6c"
"\x17\x6b\xf6\x1c\xf0\xf3\x6e\xf8"
"\x50\x38\x1f\x71\x49\xb6\x57\xd6"
"\x8f\xcb\x8d\x6b\xe3\xa6\x29\x90"
"\xfe\x2a\x62\x82\xae\x6d\x8b\xf6"
"\xad\x1e\x9e\x20\x5f\x38\xbe\x04"
"\xda\x10\x8e\xed\xa2\xa4\x87\xab"
"\xda\x6b\xb4\x0c\x75\xba\xd3\x7c"
"\xc9\xac\x42\x31\x95\x7c\xc9\x04"
"\xeb\xd5\x6e\x32\x69\x8a\xdb\xa6"
"\x15\xd7\x3f\x4f\x2f\x66\x69\x03"
"\x9c\x1f\x54\x0f\xde\x1f\xf3\x65"
"\x4c\x96\x12\xed\x7c\x92\x03\x01"
"\x6f\xbc\x35\x93\xac\xf1\x27\xf1"
"\xb4\x96\x82\x5a\x5f\xb0\xa0\x50"
"\x89\xa4\x8e\x66\x44\x85\xcc\xfd"
"\x33\x14\x70\xe3\x96\xb2\xc3\xd3"
"\xbb\x54\x5a\x1a\xf9\x74\xa2\xc5"
"\x2d\x64\x75\xdd\xb4\x54\xe6\x74"
"\x8c\xd3\x9d\x9e\x86\xab\x51\x53"
"\xb7\x93\x3e\x6f\xd0\x4e\x2c\x40"
"\xf6\xa8\x2e\x3e\x9d\xf4\x66\xa5"
"\x76\x12\x73\x44\x1a\x56\xd7\x72"
"\x88\xcd\x21\x8c\x4c\x0f\xfe\xda"
"\x95\xe0\x3a\xa6\xa5\x84\x46\xcd"
"\xd5\x3e\x9d\x3a\xe2\x67\xe6\x60"
"\x1a\xe2\x70\x85\x58\xc2\x1b\x09"
"\xe1\xd7\x2c\xca\xad\xa8\x8f\xf9"
"\xac\xb3\x0e\xdb\xca\x2e\xe2\xb8"
"\x51\x71\xd9\x3c\x6c\xf1\x56\xf8"
"\xea\x9c\xf1\xfb\x0c\xe6\xb7\x10"
"\x1c\xf8\xa9\x7c\xe8\x53\x35\xc1"
"\x90\x3e\x76\x4a\x74\xa4\x21\x2c"
"\xf6\x2c\x4e\x0f\x94\x3a\x88\x2e"
"\x41\x09\x6a\x33\x7d\xf6\xdd\x3f"
"\x8d\x23\x31\x74\x84\xeb\x88\x6e"
"\xcc\xb9\xbc\x22\x83\x19\x07\x22"
"\xa5\x2d\xdf\xa5\xf3\x80\x85\x78"
"\x84\x39\x6a\x6d\x6a\x99\x4f\xa5"
"\x15\xfe\x46\xb0\xe4\x6c\xa5\x41"
"\x3c\xce\x8f\x42\x60\x71\xa7\x75"
"\x08\x40\x65\x8a\x82\xbf\xf5\x43"
"\x71\x96\xa9\x4d\x44\x8a\x20\xbe"
"\xfa\x4d\xbb\xc0\x7d\x31\x96\x65"
"\xe7\x75\xe5\x3e\xfd\x92\x3b\xc9"
"\x55\xbb\x16\x7e\xf7\xc2\x8c\xa4"
"\x40\x1d\xe5\xef\x0e\xdf\xe4\x9a"
"\x62\x73\x65\xfd\x46\x63\x25\x3d"
"\x2b\xaf\xe5\x64\xfe\xa5\x5c\xcf"
"\x24\xf3\xb4\xac\x64\xba\xdf\x4b"
"\xc6\x96\x7d\x81\x2d\x8d\x97\xf7"
"\xc5\x68\x77\x84\x32\x2b\xcc\x85"
"\x74\x96\xf0\x12\x77\x61\xb9\xeb"
"\x71\xaa\x82\xcb\x1c\xdb\x89\xc8"
"\xc6\xb5\xe3\x5c\x7d\x39\x07\x24"
"\xda\x39\x87\x45\xc0\x2b\xbb\x01"
"\xac\xbc\x2a\x5c\x7f\xfc\xe8\xce"
"\x6d\x9c\x6f\xed\xd3\xc1\xa1\xd6"
"\xc5\x55\xa9\x66\x2f\xe1\xc8\x32"
"\xa6\x5d\xa4\x3a\x98\x73\xe8\x45"
"\xa4\xc7\xa8\xb4\xf6\x13\x03\xf6"
"\xe9\x2e\xc4\x29\x0f\x84\xdb\xc4"
"\x21\xc4\xc2\x75\x67\x89\x37\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xe3\x5a\x38\x0f\x4d\x92\x3a\x74"
"\x15\xb1\x50\x8c\x9a\xd8\x99\x1d"
"\x82\xec\xf1\x5f\x03\x6d\x02\x58"
"\x90\x67\xfc\xdd\x8d\xe1\x38\x08"
"\x7b\xc9\x9b\x4b\x04\x09\x50\x15"
"\xce\xab\xda\x33\x30\x20\x12\xfa"
"\x83\xc4\xa6\x9a\x2e\x7d\x90\xd9"
"\xa6\xa6\x67\x43\xb4\xa7\xa8\x5c"
"\xbb\x6a\x49\x2b\x8b\xf8\xd0\x22"
"\xe5\x9e\xba\xe8\x8c\x67\xb8\x5b"
"\x60\xbc\xf5\xa4\x95\x4e\x66\xe5"
"\x6d\x8e\xa9\xf6\x65\x2e\x04\xf5"
"\xba\xb5\xdb\x88\xc2\xf6\x7a\x4b"
"\x89\x58\x7c\x9a\xae\x26\xe8\xb7"
"\xb7\x28\xcc\xd6\xcc\xa5\x98\x4d"
"\xb9\x91\xcb\xb4\xe4\x8b\x96\x47"
"\x5f\x03\x8b\xdd\x94\xd1\xee\x12"
"\xa7\x83\x80\xf2\xc1\x15\x74\x4f"
"\x49\xf9\xb0\x7e\x6f\xdc\x73\x2f"
"\xe2\xcf\xe0\x1b\x34\xa5\xa0\x52"
"\xfb\x3c\x5d\x85\x91\xe6\x6d\x98"
"\x04\xd6\xdd\x4c\x00\x64\xd9\x54"
"\x5c\x3c\x08\x1d\x4c\x06\x9f\xb8"
"\x1c\x4d\x8d\xdc\xa4\x3c\xb9\x3b"
"\x9e\x85\xce\xc3\xa8\x4a\x0c\xd9"
"\x04\xc3\x6f\x17\x66\xa9\x1f\x59"
"\xd9\xe2\x19\x36\xa3\x88\xb8\x0b"
"\x0f\x4a\x4d\xf8\xc8\x6f\xd5\x43"
"\xeb\xa0\xab\x1f\x61\xc0\x06\xeb"
"\x93\xb7\xb8\x6f\x0d\xbd\x07\x49"
"\xb3\xac\x5d\xcf\x31\xa0\x27\x26"
"\x21\xbe\x94\x2e\x19\xea\xf4\xee"
"\xb5\x13\x89\xf7\x94\x0b\xef\x59"
"\x44\xc5\x78\x8b\x3c\x3b\x71\x20"
"\xf9\x35\x0c\x70\x74\xdc\x5b\xc2"
"\xb4\x11\x0e\x2c\x61\xa1\x52\x46"
"\x18\x11\x16\xc6\x86\x44\xa7\xaf"
"\xd5\x0c\x7d\xa6\x9e\x25\x2d\x1b"
"\x9a\x8f\x0f\xf8\x6a\x61\xa0\xea"
"\x3f\x0e\x90\xd6\x8f\x83\x30\x64"
"\xb5\x51\x2d\x08\x3c\xcd\x99\x36"
"\x96\xd4\xb1\xb5\x48\x30\xca\x48"
"\xf7\x11\xa8\xf5\x97\x8a\x6a\x6d"
"\x12\x33\x2f\xc0\xe8\xda\xec\x8a"
"\xe1\x88\x72\x63\xde\x20\xa3\xe1"
"\x8e\xac\x84\x37\x35\xf5\xf7\x3f"
"\x00\x02\x0e\xe4\xc1\x53\x68\x3f"
"\xaa\xd5\xac\x52\x3d\x20\x2f\x4d"
"\x7c\x83\xd0\xbd\xaa\x97\x35\x36"
"\x98\x88\x59\x5d\xe7\x24\xe3\x90"
"\x9d\x30\x47\xa7\xc3\x60\x35\xf4"
"\xd5\xdb\x0e\x4d\x44\xc1\x81\x8b"
"\xfd\xbd\xc3\x2b\xba\x68\xfe\x8d"
"\x49\x5a\x3c\x8a\xa3\x01\xae\x25"
"\x42\xab\xd2\x87\x1b\x35\xd6\xd2"
"\xd7\x70\x1c\x1f\x72\xd1\xe1\x39"
"\x1c\x58\xa2\xb4\xd0\x78\x55\x72"
"\x76\x59\xea\xd9\xd7\x6e\x63\x8b"
"\xcc\x9b\xa7\x74\x89\xfc\xa3\x68"
"\x86\x28\xd1\xbb\x54\x8d\x66\xad"
"\x2a\x92\xf9\x4e\x04\x3d\xae\xfd"
"\x1b\x2b\x7f\xc3\x2f\x1a\x78\x0a"
"\x5c\xc6\x84\xfe\x7c\xcb\x26\xfd"
"\xd9\x51\x0f\xd7\x94\x2f\xc5\xa7",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add lrw(twofish) test vectors Add test vectors for lrw(twofish). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:50 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec serpent_xts_tv_template[] = {
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
/* Generated from AES-XTS test vectors */
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
{
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xe1\x08\xb8\x1d\x2c\xf5\x33\x64"
"\xc8\x12\x04\xc7\xb3\x70\xe8\xc4"
"\x6a\x31\xc5\xf3\x00\xca\xb9\x16"
"\xde\xe2\x77\x66\xf7\xfe\x62\x08",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x1a\x0a\x09\x5f\xcd\x07\x07\x98"
"\x41\x86\x12\xaf\xb3\xd7\x68\x13"
"\xed\x81\xcd\x06\x87\x43\x1a\xbb"
"\x13\x3d\xd6\x1e\x2b\xe1\x77\xbe",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xf9\x9b\x28\xb8\x5c\xaf\x8c\x61"
"\xb6\x1c\x81\x8f\x2c\x87\x60\x89"
"\x0d\x8d\x7a\xe8\x60\x48\xcc\x86"
"\xc1\x68\x45\xaa\x00\xe9\x24\xc5",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xfe\x47\x4a\xc8\x60\x7e\xb4\x8b"
"\x0d\x10\xf4\xb0\x0d\xba\xf8\x53"
"\x65\x6e\x38\x4b\xdb\xaa\xb1\x9e"
"\x28\xca\xb0\x22\xb3\x85\x75\xf4"
"\x00\x5c\x75\x14\x06\xd6\x25\x82"
"\xe6\xcb\x08\xf7\x29\x90\x23\x8e"
"\xa4\x68\x57\xe4\xf0\xd8\x32\xf3"
"\x80\x51\x67\xb5\x0b\x85\x69\xe8"
"\x19\xfe\xc4\xc7\x3e\xea\x90\xd3"
"\x8f\xa3\xf2\x0a\xac\x17\x4b\xa0"
"\x63\x5a\x16\x0f\xf0\xce\x66\x1f"
"\x2c\x21\x07\xf1\xa4\x03\xa3\x44"
"\x41\x61\x87\x5d\x6b\xb3\xef\xd4"
"\xfc\xaa\x32\x7e\x55\x58\x04\x41"
"\xc9\x07\x33\xc6\xa2\x68\xd6\x5a"
"\x55\x79\x4b\x6f\xcf\x89\xb9\x19"
"\xe5\x54\x13\x15\xb2\x1a\xfa\x15"
"\xc2\xf0\x06\x59\xfa\xa0\x25\x05"
"\x58\xfa\x43\x91\x16\x85\x40\xbb"
"\x0d\x34\x4d\xc5\x1e\x20\xd5\x08"
"\xcd\x22\x22\x41\x11\x9f\x6c\x7c"
"\x8d\x57\xc9\xba\x57\xe8\x2c\xf7"
"\xa0\x42\xa8\xde\xfc\xa3\xca\x98"
"\x4b\x43\xb1\xce\x4b\xbf\x01\x67"
"\x6e\x29\x60\xbd\x10\x14\x84\x82"
"\x83\x82\x0c\x63\x73\x92\x02\x7c"
"\x55\x37\x20\x80\x17\x51\xc8\xbc"
"\x46\x02\xcb\x38\x07\x6d\xe2\x85"
"\xaa\x29\xaf\x24\x58\x0d\xf0\x75"
"\x08\x0a\xa5\x34\x25\x16\xf3\x74"
"\xa7\x0b\x97\xbe\xc1\xa9\xdc\x29"
"\x1a\x0a\x56\xc1\x1a\x91\x97\x8c"
"\x0b\xc7\x16\xed\x5a\x22\xa6\x2e"
"\x8c\x2b\x4f\x54\x76\x47\x53\x8e"
"\xe8\x00\xec\x92\xb9\x55\xe6\xa2"
"\xf3\xe2\x4f\x6a\x66\x60\xd0\x87"
"\xe6\xd1\xcc\xe3\x6a\xc5\x2d\x21"
"\xcc\x9d\x6a\xb6\x75\xaa\xe2\x19"
"\x21\x9f\xa1\x5e\x4c\xfd\x72\xf9"
"\x94\x4e\x63\xc7\xae\xfc\xed\x47"
"\xe2\xfe\x7a\x63\x77\xfe\x97\x82"
"\xb1\x10\x6e\x36\x1d\xe1\xc4\x80"
"\xec\x69\x41\xec\xa7\x8a\xe0\x2f"
"\xe3\x49\x26\xa2\x41\xb2\x08\x0f"
"\x28\xb4\xa7\x39\xa1\x99\x2d\x1e"
"\x43\x42\x35\xd0\xcf\xec\x77\x67"
"\xb2\x3b\x9e\x1c\x35\xde\x4f\x5e"
"\x73\x3f\x5d\x6f\x07\x4b\x2e\x50"
"\xab\x6c\x6b\xff\xea\x00\x67\xaa"
"\x0e\x82\x32\xdd\x3d\xb5\xe5\x76"
"\x2b\x77\x3f\xbe\x12\x75\xfb\x92"
"\xc6\x89\x67\x4d\xca\xf7\xd4\x50"
"\xc0\x74\x47\xcc\xd9\x0a\xd4\xc6"
"\x3b\x17\x2e\xe3\x35\xbb\x53\xb5"
"\x86\xad\x51\xcc\xd5\x96\xb8\xdc"
"\x03\x57\xe6\x98\x52\x2f\x61\x62"
"\xc4\x5c\x9c\x36\x71\x07\xfb\x94"
"\xe3\x02\xc4\x2b\x08\x75\xc7\x35"
"\xfb\x2e\x88\x7b\xbb\x67\x00\xe1"
"\xc9\xdd\x99\xb2\x13\x53\x1a\x4e"
"\x76\x87\x19\x04\x1a\x2f\x38\x3e"
"\xef\x91\x64\x1d\x18\x07\x4e\x31"
"\x88\x21\x7c\xb0\xa5\x12\x4c\x3c"
"\xb0\x20\xbd\xda\xdf\xf9\x7c\xdd",
.len = 512,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x2b\xc9\xb4\x6b\x10\x94\xa9\x32"
"\xaa\xb0\x20\xc6\x44\x3d\x74\x1f"
"\x75\x01\xa7\xf6\xf5\xf7\x62\x1b"
"\x80\x1b\x82\xcb\x01\x59\x91\x7f"
"\x80\x3a\x98\xf0\xd2\xca\xc4\xc3"
"\x34\xfd\xe6\x11\xf9\x33\x45\x12"
"\x48\xc5\x8c\x25\xf1\xc5\xc5\x23"
"\xd3\x44\xb4\x73\xd5\x04\xc0\xb7"
"\xca\x2f\xf5\xcd\xc5\xb4\xdd\xb0"
"\xf4\x60\xe8\xfb\xc6\x9c\xc5\x78"
"\xcd\xec\x7d\xdc\x19\x9c\x72\x64"
"\x63\x0b\x38\x2e\x76\xdd\x2d\x36"
"\x49\xb0\x1d\xea\x78\x9e\x00\xca"
"\x20\xcc\x1b\x1e\x98\x74\xab\xed"
"\x79\xf7\xd0\x6c\xd8\x93\x80\x29"
"\xac\xa5\x5e\x34\xa9\xab\xa0\x55"
"\x9a\xea\xaa\x95\x4d\x7b\xfe\x46"
"\x26\x8a\xfd\x88\xa2\xa8\xa6\xae"
"\x25\x42\x17\xbf\x76\x8f\x1c\x3d"
"\xec\x9a\xda\x64\x96\xb5\x61\xff"
"\x99\xeb\x12\x96\x85\x82\x9d\xd5"
"\x81\x85\x14\xa8\x59\xac\x8c\x94"
"\xbb\x3b\x85\x2b\xdf\xb3\x0c\xba"
"\x82\xc6\x4d\xca\x86\xea\x53\x28"
"\x4c\xe0\x4e\x31\xe3\x73\x2f\x79"
"\x9d\x42\xe1\x03\xe3\x8b\xc4\xff"
"\x05\xca\x81\x7b\xda\xa2\xde\x63"
"\x3a\x10\xbe\xc2\xac\x32\xc4\x05"
"\x47\x7e\xef\x67\xe2\x5f\x5b\xae"
"\xed\xf1\x70\x34\x16\x9a\x07\x7b"
"\xf2\x25\x2b\xb0\xf8\x3c\x15\x9a"
"\xa6\x59\x55\x5f\xc1\xf4\x1e\xcd"
"\x93\x1f\x06\xba\xd4\x9a\x22\x69"
"\xfa\x8e\x95\x0d\xf3\x23\x59\x2c"
"\xfe\x00\xba\xf0\x0e\xbc\x6d\xd6"
"\x62\xf0\x7a\x0e\x83\x3e\xdb\x32"
"\xfd\x43\x7d\xda\x42\x51\x87\x43"
"\x9d\xf9\xef\xf4\x30\x97\xf8\x09"
"\x88\xfc\x3f\x93\x70\xc1\x4a\xec"
"\x27\x5f\x11\xac\x71\xc7\x48\x46"
"\x2f\xf9\xdf\x8d\x9f\xf7\x2e\x56"
"\x0d\x4e\xb0\x32\x76\xce\x86\x81"
"\xcd\xdf\xe4\x00\xbf\xfd\x5f\x24"
"\xaf\xf7\x9a\xde\xff\x18\xac\x14"
"\x90\xc5\x01\x39\x34\x0f\x24\xf3"
"\x13\x2f\x5e\x4f\x30\x9a\x36\x40"
"\xec\xea\xbc\xcd\x9e\x0e\x5b\x23"
"\x50\x88\x97\x40\x69\xb1\x37\xf5"
"\xc3\x15\xf9\x3f\xb7\x79\x64\xe8"
"\x7b\x10\x20\xb9\x2b\x46\x83\x5b"
"\xd8\x39\xfc\xe4\xfa\x88\x52\xf2"
"\x72\xb0\x97\x4e\x89\xb3\x48\x00"
"\xc1\x16\x73\x50\x77\xba\xa6\x65"
"\x20\x2d\xb0\x02\x27\x89\xda\x99"
"\x45\xfb\xe9\xd3\x1d\x39\x2f\xd6"
"\x2a\xda\x09\x12\x11\xaf\xe6\x57"
"\x01\x04\x8a\xff\x86\x8b\xac\xf8"
"\xee\xe4\x1c\x98\x5b\xcf\x6b\x76"
"\xa3\x0e\x33\x74\x40\x18\x39\x72"
"\x66\x50\x31\xfd\x70\xdf\xe8\x51"
"\x96\x21\x36\xb2\x9b\xfa\x85\xd1"
"\x30\x05\xc8\x92\x98\x80\xff\x7a"
"\xaf\x43\x0b\xc5\x20\x41\x92\x20"
"\xd4\xa0\x91\x98\x11\x5f\x4d\xb1",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* SM4 test vector taken from the draft RFC
* https://tools.ietf.org/html/draft-crypto-sm4-00#ref-GBT.32907-2016
*/
static const struct cipher_testvec sm4_tv_template[] = {
{ /* SM4 Appendix A: Example Calculations. Example 1. */
.key = "\x01\x23\x45\x67\x89\xAB\xCD\xEF"
"\xFE\xDC\xBA\x98\x76\x54\x32\x10",
.klen = 16,
.ptext = "\x01\x23\x45\x67\x89\xAB\xCD\xEF"
"\xFE\xDC\xBA\x98\x76\x54\x32\x10",
.ctext = "\x68\x1E\xDF\x34\xD2\x06\x96\x5E"
"\x86\xB3\xE9\x4F\x53\x6E\x42\x46",
.len = 16,
}, { /*
* SM4 Appendix A: Example Calculations.
* Last 10 iterations of Example 2.
*/
.key = "\x01\x23\x45\x67\x89\xAB\xCD\xEF"
"\xFE\xDC\xBA\x98\x76\x54\x32\x10",
.klen = 16,
.ptext = "\x99\x4a\xc3\xe7\xc3\x57\x89\x6a"
"\x81\xfc\xa8\xe\x38\x3e\xef\x80"
"\xb1\x98\xf2\xde\x3f\x4b\xae\xd1"
"\xf0\xf1\x30\x4c\x1\x27\x5a\x8f"
"\x45\xe1\x39\xb7\xae\xff\x1f\x27"
"\xad\x57\x15\xab\x31\x5d\xc\xef"
"\x8c\xc8\x80\xbd\x11\x98\xf3\x7b"
"\xa2\xdd\x14\x20\xf9\xe8\xbb\x82"
"\xf7\x32\xca\x4b\xa8\xf7\xb3\x4d"
"\x27\xd1\xcd\xe6\xb6\x65\x5a\x23"
"\xc2\xf3\x54\x84\x53\xe3\xb9\x20"
"\xa5\x37\x0\xbe\xe7\x7b\x48\xfb"
"\x21\x3d\x9e\x48\x1d\x9e\xf5\xbf"
"\x77\xd5\xb4\x4a\x53\x71\x94\x7a"
"\x88\xa6\x6e\x6\x93\xca\x43\xa5"
"\xc4\xf6\xcd\x53\x4b\x7b\x8e\xfe"
"\xb4\x28\x7c\x42\x29\x32\x5d\x88"
"\xed\xce\x0\x19\xe\x16\x2\x6e"
"\x87\xff\x2c\xac\xe8\xe7\xe9\xbf"
"\x31\x51\xec\x47\xc3\x51\x83\xc1",
.ctext = "\xb1\x98\xf2\xde\x3f\x4b\xae\xd1"
"\xf0\xf1\x30\x4c\x1\x27\x5a\x8f"
"\x45\xe1\x39\xb7\xae\xff\x1f\x27"
"\xad\x57\x15\xab\x31\x5d\xc\xef"
"\x8c\xc8\x80\xbd\x11\x98\xf3\x7b"
"\xa2\xdd\x14\x20\xf9\xe8\xbb\x82"
"\xf7\x32\xca\x4b\xa8\xf7\xb3\x4d"
"\x27\xd1\xcd\xe6\xb6\x65\x5a\x23"
"\xc2\xf3\x54\x84\x53\xe3\xb9\x20"
"\xa5\x37\x0\xbe\xe7\x7b\x48\xfb"
"\x21\x3d\x9e\x48\x1d\x9e\xf5\xbf"
"\x77\xd5\xb4\x4a\x53\x71\x94\x7a"
"\x88\xa6\x6e\x6\x93\xca\x43\xa5"
"\xc4\xf6\xcd\x53\x4b\x7b\x8e\xfe"
"\xb4\x28\x7c\x42\x29\x32\x5d\x88"
"\xed\xce\x0\x19\xe\x16\x2\x6e"
"\x87\xff\x2c\xac\xe8\xe7\xe9\xbf"
"\x31\x51\xec\x47\xc3\x51\x83\xc1"
"\x59\x52\x98\xc7\xc6\xfd\x27\x1f"
"\x4\x2\xf8\x4\xc3\x3d\x3f\x66",
.len = 160
}
};
/*
* Speck test vectors taken from the original paper:
* "The Simon and Speck Families of Lightweight Block Ciphers"
* https://eprint.iacr.org/2013/404.pdf
*
* Note that the paper does not make byte and word order clear. But it was
* confirmed with the authors that the intended orders are little endian byte
* order and (y, x) word order. Equivalently, the printed test vectors, when
* looking at only the bytes (ignoring the whitespace that divides them into
* words), are backwards: the left-most byte is actually the one with the
* highest memory address, while the right-most byte is actually the one with
* the lowest memory address.
*/
static const struct cipher_testvec speck128_tv_template[] = {
{ /* Speck128/128 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.klen = 16,
.ptext = "\x20\x6d\x61\x64\x65\x20\x69\x74"
"\x20\x65\x71\x75\x69\x76\x61\x6c",
.ctext = "\x18\x0d\x57\x5c\xdf\xfe\x60\x78"
"\x65\x32\x78\x79\x51\x98\x5d\xa6",
.len = 16,
}, { /* Speck128/192 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17",
.klen = 24,
.ptext = "\x65\x6e\x74\x20\x74\x6f\x20\x43"
"\x68\x69\x65\x66\x20\x48\x61\x72",
.ctext = "\x86\x18\x3c\xe0\x5d\x18\xbc\xf9"
"\x66\x55\x13\x13\x3a\xcf\xe4\x1b",
.len = 16,
}, { /* Speck128/256 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.klen = 32,
.ptext = "\x70\x6f\x6f\x6e\x65\x72\x2e\x20"
"\x49\x6e\x20\x74\x68\x6f\x73\x65",
.ctext = "\x43\x8f\x18\x9c\x8d\xb4\xee\x4e"
"\x3e\xf5\xc0\x05\x04\x01\x09\x41",
.len = 16,
},
};
/*
* Speck128-XTS test vectors, taken from the AES-XTS test vectors with the
* ciphertext recomputed with Speck128 as the cipher
*/
static const struct cipher_testvec speck128_xts_tv_template[] = {
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
{
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xbe\xa0\xe7\x03\xd7\xfe\xab\x62"
"\x3b\x99\x4a\x64\x74\x77\xac\xed"
"\xd8\xf4\xa6\xcf\xae\xb9\x07\x42"
"\x51\xd9\xb6\x1d\xe0\x5e\xbc\x54",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xfb\x53\x81\x75\x6f\x9f\x34\xad"
"\x7e\x01\xed\x7b\xcc\xda\x4e\x4a"
"\xd4\x84\xa4\x53\xd5\x88\x73\x1b"
"\xfd\xcb\xae\x0d\xf3\x04\xee\xe6",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x21\x52\x84\x15\xd1\xf7\x21\x55"
"\xd9\x75\x4a\xd3\xc5\xdb\x9f\x7d"
"\xda\x63\xb2\xf1\x82\xb0\x89\x59"
"\x86\xd4\xaa\xaa\xdd\xff\x4f\x92",
.len = 32,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x57\xb5\xf8\x71\x6e\x6d\xdd\x82"
"\x53\xd0\xed\x2d\x30\xc1\x20\xef"
"\x70\x67\x5e\xff\x09\x70\xbb\xc1"
"\x3a\x7b\x48\x26\xd9\x0b\xf4\x48"
"\xbe\xce\xb1\xc7\xb2\x67\xc4\xa7"
"\x76\xf8\x36\x30\xb7\xb4\x9a\xd9"
"\xf5\x9d\xd0\x7b\xc1\x06\x96\x44"
"\x19\xc5\x58\x84\x63\xb9\x12\x68"
"\x68\xc7\xaa\x18\x98\xf2\x1f\x5c"
"\x39\xa6\xd8\x32\x2b\xc3\x51\xfd"
"\x74\x79\x2e\xb4\x44\xd7\x69\xc4"
"\xfc\x29\xe6\xed\x26\x1e\xa6\x9d"
"\x1c\xbe\x00\x0e\x7f\x3a\xca\xfb"
"\x6d\x13\x65\xa0\xf9\x31\x12\xe2"
"\x26\xd1\xec\x2b\x0a\x8b\x59\x99"
"\xa7\x49\xa0\x0e\x09\x33\x85\x50"
"\xc3\x23\xca\x7a\xdd\x13\x45\x5f"
"\xde\x4c\xa7\xcb\x00\x8a\x66\x6f"
"\xa2\xb6\xb1\x2e\xe1\xa0\x18\xf6"
"\xad\xf3\xbd\xeb\xc7\xef\x55\x4f"
"\x79\x91\x8d\x36\x13\x7b\xd0\x4a"
"\x6c\x39\xfb\x53\xb8\x6f\x02\x51"
"\xa5\x20\xac\x24\x1c\x73\x59\x73"
"\x58\x61\x3a\x87\x58\xb3\x20\x56"
"\x39\x06\x2b\x4d\xd3\x20\x2b\x89"
"\x3f\xa2\xf0\x96\xeb\x7f\xa4\xcd"
"\x11\xae\xbd\xcb\x3a\xb4\xd9\x91"
"\x09\x35\x71\x50\x65\xac\x92\xe3"
"\x7b\x32\xc0\x7a\xdd\xd4\xc3\x92"
"\x6f\xeb\x79\xde\x6f\xd3\x25\xc9"
"\xcd\x63\xf5\x1e\x7a\x3b\x26\x9d"
"\x77\x04\x80\xa9\xbf\x38\xb5\xbd"
"\xb8\x05\x07\xbd\xfd\xab\x7b\xf8"
"\x2a\x26\xcc\x49\x14\x6d\x55\x01"
"\x06\x94\xd8\xb2\x2d\x53\x83\x1b"
"\x8f\xd4\xdd\x57\x12\x7e\x18\xba"
"\x8e\xe2\x4d\x80\xef\x7e\x6b\x9d"
"\x24\xa9\x60\xa4\x97\x85\x86\x2a"
"\x01\x00\x09\xf1\xcb\x4a\x24\x1c"
"\xd8\xf6\xe6\x5b\xe7\x5d\xf2\xc4"
"\x97\x1c\x10\xc6\x4d\x66\x4f\x98"
"\x87\x30\xac\xd5\xea\x73\x49\x10"
"\x80\xea\xe5\x5f\x4d\x5f\x03\x33"
"\x66\x02\x35\x3d\x60\x06\x36\x4f"
"\x14\x1c\xd8\x07\x1f\x78\xd0\xf8"
"\x4f\x6c\x62\x7c\x15\xa5\x7c\x28"
"\x7c\xcc\xeb\x1f\xd1\x07\x90\x93"
"\x7e\xc2\xa8\x3a\x80\xc0\xf5\x30"
"\xcc\x75\xcf\x16\x26\xa9\x26\x3b"
"\xe7\x68\x2f\x15\x21\x5b\xe4\x00"
"\xbd\x48\x50\xcd\x75\x70\xc4\x62"
"\xbb\x41\xfb\x89\x4a\x88\x3b\x3b"
"\x51\x66\x02\x69\x04\x97\x36\xd4"
"\x75\xae\x0b\xa3\x42\xf8\xca\x79"
"\x8f\x93\xe9\xcc\x38\xbd\xd6\xd2"
"\xf9\x70\x4e\xc3\x6a\x8e\x25\xbd"
"\xea\x15\x5a\xa0\x85\x7e\x81\x0d"
"\x03\xe7\x05\x39\xf5\x05\x26\xee"
"\xec\xaa\x1f\x3d\xc9\x98\x76\x01"
"\x2c\xf4\xfc\xa3\x88\x77\x38\xc4"
"\x50\x65\x50\x6d\x04\x1f\xdf\x5a"
"\xaa\xf2\x01\xa9\xc1\x8d\xee\xca"
"\x47\x26\xef\x39\xb8\xb4\xf2\xd1"
"\xd6\xbb\x1b\x2a\xc1\x34\x14\xcf",
.len = 512,
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(twofish) test vectors Add test vectors for xts(twofish). These are generated from xts(twofish) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:33 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc5\x85\x2a\x4b\x73\xe4\xf6\xf1"
"\x7e\xf9\xf6\xe9\xa3\x73\x36\xcb"
"\xaa\xb6\x22\xb0\x24\x6e\x3d\x73"
"\x92\x99\xde\xd3\x76\xed\xcd\x63"
"\x64\x3a\x22\x57\xc1\x43\x49\xd4"
"\x79\x36\x31\x19\x62\xae\x10\x7e"
"\x7d\xcf\x7a\xe2\x6b\xce\x27\xfa"
"\xdc\x3d\xd9\x83\xd3\x42\x4c\xe0"
"\x1b\xd6\x1d\x1a\x6f\xd2\x03\x00"
"\xfc\x81\x99\x8a\x14\x62\xf5\x7e"
"\x0d\xe7\x12\xe8\x17\x9d\x0b\xec"
"\xe2\xf7\xc9\xa7\x63\xd1\x79\xb6"
"\x62\x62\x37\xfe\x0a\x4c\x4a\x37"
"\x70\xc7\x5e\x96\x5f\xbc\x8e\x9e"
"\x85\x3c\x4f\x26\x64\x85\xbc\x68"
"\xb0\xe0\x86\x5e\x26\x41\xce\x11"
"\x50\xda\x97\x14\xe9\x9e\xc7\x6d"
"\x3b\xdc\x43\xde\x2b\x27\x69\x7d"
"\xfc\xb0\x28\xbd\x8f\xb1\xc6\x31"
"\x14\x4d\xf0\x74\x37\xfd\x07\x25"
"\x96\x55\xe5\xfc\x9e\x27\x2a\x74"
"\x1b\x83\x4d\x15\x83\xac\x57\xa0"
"\xac\xa5\xd0\x38\xef\x19\x56\x53"
"\x25\x4b\xfc\xce\x04\x23\xe5\x6b"
"\xf6\xc6\x6c\x32\x0b\xb3\x12\xc5"
"\xed\x22\x34\x1c\x5d\xed\x17\x06"
"\x36\xa3\xe6\x77\xb9\x97\x46\xb8"
"\xe9\x3f\x7e\xc7\xbc\x13\x5c\xdc"
"\x6e\x3f\x04\x5e\xd1\x59\xa5\x82"
"\x35\x91\x3d\x1b\xe4\x97\x9f\x92"
"\x1c\x5e\x5f\x6f\x41\xd4\x62\xa1"
"\x8d\x39\xfc\x42\xfb\x38\x80\xb9"
"\x0a\xe3\xcc\x6a\x93\xd9\x7a\xb1"
"\xe9\x69\xaf\x0a\x6b\x75\x38\xa7"
"\xa1\xbf\xf7\xda\x95\x93\x4b\x78"
"\x19\xf5\x94\xf9\xd2\x00\x33\x37"
"\xcf\xf5\x9e\x9c\xf3\xcc\xa6\xee"
"\x42\xb2\x9e\x2c\x5f\x48\x23\x26"
"\x15\x25\x17\x03\x3d\xfe\x2c\xfc"
"\xeb\xba\xda\xe0\x00\x05\xb6\xa6"
"\x07\xb3\xe8\x36\x5b\xec\x5b\xbf"
"\xd6\x5b\x00\x74\xc6\x97\xf1\x6a"
"\x49\xa1\xc3\xfa\x10\x52\xb9\x14"
"\xad\xb7\x73\xf8\x78\x12\xc8\x59"
"\x17\x80\x4c\x57\x39\xf1\x6d\x80"
"\x25\x77\x0f\x5e\x7d\xf0\xaf\x21"
"\xec\xce\xb7\xc8\x02\x8a\xed\x53"
"\x2c\x25\x68\x2e\x1f\x85\x5e\x67"
"\xd1\x07\x7a\x3a\x89\x08\xe0\x34"
"\xdc\xdb\x26\xb4\x6b\x77\xfc\x40"
"\x31\x15\x72\xa0\xf0\x73\xd9\x3b"
"\xd5\xdb\xfe\xfc\x8f\xa9\x44\xa2"
"\x09\x9f\xc6\x33\xe5\xe2\x88\xe8"
"\xf3\xf0\x1a\xf4\xce\x12\x0f\xd6"
"\xf7\x36\xe6\xa4\xf4\x7a\x10\x58"
"\xcc\x1f\x48\x49\x65\x47\x75\xe9"
"\x28\xe1\x65\x7b\xf2\xc4\xb5\x07"
"\xf2\xec\x76\xd8\x8f\x09\xf3\x16"
"\xa1\x51\x89\x3b\xeb\x96\x42\xac"
"\x65\xe0\x67\x63\x29\xdc\xb4\x7d"
"\xf2\x41\x51\x6a\xcb\xde\x3c\xfb"
"\x66\x8d\x13\xca\xe0\x59\x2a\x00"
"\xc9\x53\x4c\xe6\x9e\xe2\x73\xd5"
"\x67\x19\xb2\xbd\x9a\x63\xd7\x5c",
.len = 512,
.also_non_np = 1,
.np = 3,
.tap = { 512 - 20, 4, 16 },
}
};
static const struct cipher_testvec speck64_tv_template[] = {
{ /* Speck64/96 */
.key = "\x00\x01\x02\x03\x08\x09\x0a\x0b"
"\x10\x11\x12\x13",
.klen = 12,
.ptext = "\x65\x61\x6e\x73\x20\x46\x61\x74",
.ctext = "\x6c\x94\x75\x41\xec\x52\x79\x9f",
.len = 8,
}, { /* Speck64/128 */
.key = "\x00\x01\x02\x03\x08\x09\x0a\x0b"
"\x10\x11\x12\x13\x18\x19\x1a\x1b",
.klen = 16,
.ptext = "\x2d\x43\x75\x74\x74\x65\x72\x3b",
.ctext = "\x8b\x02\x4e\x45\x48\xa5\x6f\x8c",
.len = 8,
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* Speck64-XTS test vectors, taken from the AES-XTS test vectors with the
* ciphertext recomputed with Speck64 as the cipher, and key lengths adjusted
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec speck64_xts_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 24,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x84\xaf\x54\x07\x19\xd4\x7c\xa6"
"\xe4\xfe\xdf\xc4\x1f\x34\xc3\xc2"
"\x80\xf5\x72\xe7\xcd\xf0\x99\x22"
"\x35\xa7\x2f\x06\xef\xdc\x51\xaa",
.len = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 24,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\x12\x56\x73\xcd\x15\x87\xa8\x59"
"\xcf\x84\xae\xd9\x1c\x66\xd6\x9f"
"\xb3\x12\x69\x7e\x36\xeb\x52\xff"
"\x62\xdd\xba\x90\xb3\xe1\xee\x99",
.len = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 24,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\x15\x1b\xe4\x2c\xa2\x5a\x2d\x2c"
"\x27\x36\xc0\xbf\x5d\xea\x36\x37"
"\x2d\x1a\x88\xbc\x66\xb5\xd0\x0b"
"\xa1\xbc\x19\xb2\x0f\x3b\x75\x34",
.len = 32,
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93",
.klen = 24,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ctext = "\xaf\xa1\x81\xa6\x32\xbb\x15\x8e"
"\xf8\x95\x2e\xd3\xe6\xee\x7e\x09"
"\x0c\x1a\xf5\x02\x97\x8b\xe3\xb3"
"\x11\xc7\x39\x96\xd0\x95\xf4\x56"
"\xf4\xdd\x03\x38\x01\x44\x2c\xcf"
"\x88\xae\x8e\x3c\xcd\xe7\xaa\x66"
"\xfe\x3d\xc6\xfb\x01\x23\x51\x43"
"\xd5\xd2\x13\x86\x94\x34\xe9\x62"
"\xf9\x89\xe3\xd1\x7b\xbe\xf8\xef"
"\x76\x35\x04\x3f\xdb\x23\x9d\x0b"
"\x85\x42\xb9\x02\xd6\xcc\xdb\x96"
"\xa7\x6b\x27\xb6\xd4\x45\x8f\x7d"
"\xae\xd2\x04\xd5\xda\xc1\x7e\x24"
"\x8c\x73\xbe\x48\x7e\xcf\x65\x28"
"\x29\xe5\xbe\x54\x30\xcb\x46\x95"
"\x4f\x2e\x8a\x36\xc8\x27\xc5\xbe"
"\xd0\x1a\xaf\xab\x26\xcd\x9e\x69"
"\xa1\x09\x95\x71\x26\xe9\xc4\xdf"
"\xe6\x31\xc3\x46\xda\xaf\x0b\x41"
"\x1f\xab\xb1\x8e\xd6\xfc\x0b\xb3"
"\x82\xc0\x37\x27\xfc\x91\xa7\x05"
"\xfb\xc5\xdc\x2b\x74\x96\x48\x43"
"\x5d\x9c\x19\x0f\x60\x63\x3a\x1f"
"\x6f\xf0\x03\xbe\x4d\xfd\xc8\x4a"
"\xc6\xa4\x81\x6d\xc3\x12\x2a\x5c"
"\x07\xff\xf3\x72\x74\x48\xb5\x40"
"\x50\xb5\xdd\x90\x43\x31\x18\x15"
"\x7b\xf2\xa6\xdb\x83\xc8\x4b\x4a"
"\x29\x93\x90\x8b\xda\x07\xf0\x35"
"\x6d\x90\x88\x09\x4e\x83\xf5\x5b"
"\x94\x12\xbb\x33\x27\x1d\x3f\x23"
"\x51\xa8\x7c\x07\xa2\xae\x77\xa6"
"\x50\xfd\xcc\xc0\x4f\x80\x7a\x9f"
"\x66\xdd\xcd\x75\x24\x8b\x33\xf7"
"\x20\xdb\x83\x9b\x4f\x11\x63\x6e"
"\xcf\x37\xef\xc9\x11\x01\x5c\x45"
"\x32\x99\x7c\x3c\x9e\x42\x89\xe3"
"\x70\x6d\x15\x9f\xb1\xe6\xb6\x05"
"\xfe\x0c\xb9\x49\x2d\x90\x6d\xcc"
"\x5d\x3f\xc1\xfe\x89\x0a\x2e\x2d"
"\xa0\xa8\x89\x3b\x73\x39\xa5\x94"
"\x4c\xa4\xa6\xbb\xa7\x14\x46\x89"
"\x10\xff\xaf\xef\xca\xdd\x4f\x80"
"\xb3\xdf\x3b\xab\xd4\xe5\x5a\xc7"
"\x33\xca\x00\x8b\x8b\x3f\xea\xec"
"\x68\x8a\xc2\x6d\xfd\xd4\x67\x0f"
"\x22\x31\xe1\x0e\xfe\x5a\x04\xd5"
"\x64\xa3\xf1\x1a\x76\x28\xcc\x35"
"\x36\xa7\x0a\x74\xf7\x1c\x44\x9b"
"\xc7\x1b\x53\x17\x02\xea\xd1\xad"
"\x13\x51\x73\xc0\xa0\xb2\x05\x32"
"\xa8\xa2\x37\x2e\xe1\x7a\x3a\x19"
"\x26\xb4\x6c\x62\x5d\xb3\x1a\x1d"
"\x59\xda\xee\x1a\x22\x18\xda\x0d"
"\x88\x0f\x55\x8b\x72\x62\xfd\xc1"
"\x69\x13\xcd\x0d\x5f\xc1\x09\x52"
"\xee\xd6\xe3\x84\x4d\xee\xf6\x88"
"\xaf\x83\xdc\x76\xf4\xc0\x93\x3f"
"\x4a\x75\x2f\xb0\x0b\x3e\xc4\x54"
"\x7d\x69\x8d\x00\x62\x77\x0d\x14"
"\xbe\x7c\xa6\x7d\xc5\x24\x4f\xf3"
"\x50\xf7\x5f\xf4\xc2\xca\x41\x97"
"\x37\xbe\x75\x74\xcd\xf0\x75\x6e"
"\x25\x23\x94\xbd\xda\x8d\xb0\xd4",
.len = 512,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27",
.klen = 32,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ctext = "\x55\xed\x71\xd3\x02\x8e\x15\x3b"
"\xc6\x71\x29\x2d\x3e\x89\x9f\x59"
"\x68\x6a\xcc\x8a\x56\x97\xf3\x95"
"\x4e\x51\x08\xda\x2a\xf8\x6f\x3c"
"\x78\x16\xea\x80\xdb\x33\x75\x94"
"\xf9\x29\xc4\x2b\x76\x75\x97\xc7"
"\xf2\x98\x2c\xf9\xff\xc8\xd5\x2b"
"\x18\xf1\xaf\xcf\x7c\xc5\x0b\xee"
"\xad\x3c\x76\x7c\xe6\x27\xa2\x2a"
"\xe4\x66\xe1\xab\xa2\x39\xfc\x7c"
"\xf5\xec\x32\x74\xa3\xb8\x03\x88"
"\x52\xfc\x2e\x56\x3f\xa1\xf0\x9f"
"\x84\x5e\x46\xed\x20\x89\xb6\x44"
"\x8d\xd0\xed\x54\x47\x16\xbe\x95"
"\x8a\xb3\x6b\x72\xc4\x32\x52\x13"
"\x1b\xb0\x82\xbe\xac\xf9\x70\xa6"
"\x44\x18\xdd\x8c\x6e\xca\x6e\x45"
"\x8f\x1e\x10\x07\x57\x25\x98\x7b"
"\x17\x8c\x78\xdd\x80\xa7\xd9\xd8"
"\x63\xaf\xb9\x67\x57\xfd\xbc\xdb"
"\x44\xe9\xc5\x65\xd1\xc7\x3b\xff"
"\x20\xa0\x80\x1a\xc3\x9a\xad\x5e"
"\x5d\x3b\xd3\x07\xd9\xf5\xfd\x3d"
"\x4a\x8b\xa8\xd2\x6e\x7a\x51\x65"
"\x6c\x8e\x95\xe0\x45\xc9\x5f\x4a"
"\x09\x3c\x3d\x71\x7f\x0c\x84\x2a"
"\xc8\x48\x52\x1a\xc2\xd5\xd6\x78"
"\x92\x1e\xa0\x90\x2e\xea\xf0\xf3"
"\xdc\x0f\xb1\xaf\x0d\x9b\x06\x2e"
"\x35\x10\x30\x82\x0d\xe7\xc5\x9b"
"\xde\x44\x18\xbd\x9f\xd1\x45\xa9"
"\x7b\x7a\x4a\xad\x35\x65\x27\xca"
"\xb2\xc3\xd4\x9b\x71\x86\x70\xee"
"\xf1\x89\x3b\x85\x4b\x5b\xaa\xaf"
"\xfc\x42\xc8\x31\x59\xbe\x16\x60"
"\x4f\xf9\xfa\x12\xea\xd0\xa7\x14"
"\xf0\x7a\xf3\xd5\x8d\xbd\x81\xef"
"\x52\x7f\x29\x51\x94\x20\x67\x3c"
"\xd1\xaf\x77\x9f\x22\x5a\x4e\x63"
"\xe7\xff\x73\x25\xd1\xdd\x96\x8a"
"\x98\x52\x6d\xf3\xac\x3e\xf2\x18"
"\x6d\xf6\x0a\x29\xa6\x34\x3d\xed"
"\xe3\x27\x0d\x9d\x0a\x02\x44\x7e"
"\x5a\x7e\x67\x0f\x0a\x9e\xd6\xad"
"\x91\xe6\x4d\x81\x8c\x5c\x59\xaa"
"\xfb\xeb\x56\x53\xd2\x7d\x4c\x81"
"\x65\x53\x0f\x41\x11\xbd\x98\x99"
"\xf9\xc6\xfa\x51\x2e\xa3\xdd\x8d"
"\x84\x98\xf9\x34\xed\x33\x2a\x1f"
"\x82\xed\xc1\x73\x98\xd3\x02\xdc"
"\xe6\xc2\x33\x1d\xa2\xb4\xca\x76"
"\x63\x51\x34\x9d\x96\x12\xae\xce"
"\x83\xc9\x76\x5e\xa4\x1b\x53\x37"
"\x17\xd5\xc0\x80\x1d\x62\xf8\x3d"
"\x54\x27\x74\xbb\x10\x86\x57\x46"
"\x68\xe1\xed\x14\xe7\x9d\xfc\x84"
"\x47\xbc\xc2\xf8\x19\x4b\x99\xcf"
"\x7a\xe9\xc4\xb8\x8c\x82\x72\x4d"
"\x7b\x4f\x38\x55\x36\x71\x64\xc1"
"\xfc\x5c\x75\x52\x33\x02\x18\xf8"
"\x17\xe1\x2b\xc2\x43\x39\xbd\x76"
"\x9b\x63\x76\x32\x2f\x19\x72\x10"
"\x9f\x21\x0c\xf1\x66\x50\x7f\xa5"
"\x0d\x1f\x46\xe0\xba\xd3\x2f\x3c",
.len = 512,
.also_non_np = 1,
.np = 3,
.tap = { 512 - 20, 4, 16 },
}
};
/* Cast6 test vectors from RFC 2612 */
static const struct cipher_testvec cast6_tv_template[] = {
{
.key = "\x23\x42\xbb\x9e\xfa\x38\x54\x2c"
"\x0a\xf7\x56\x47\xf2\x9f\x61\x5d",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = zeroed_string,
.ctext = "\xc8\x42\xa0\x89\x72\xb4\x3d\x20"
"\x83\x6c\x91\xd1\xb7\x53\x0f\x6b",
.len = 16,
}, {
.key = "\x23\x42\xbb\x9e\xfa\x38\x54\x2c"
"\xbe\xd0\xac\x83\x94\x0a\xc2\x98"
"\xba\xc7\x7a\x77\x17\x94\x28\x63",
.klen = 24,
.ptext = zeroed_string,
.ctext = "\x1b\x38\x6c\x02\x10\xdc\xad\xcb"
"\xdd\x0e\x41\xaa\x08\xa7\xa7\xe8",
.len = 16,
}, {
.key = "\x23\x42\xbb\x9e\xfa\x38\x54\x2c"
"\xbe\xd0\xac\x83\x94\x0a\xc2\x98"
"\x8d\x7c\x47\xce\x26\x49\x08\x46"
"\x1c\xc1\xb5\x13\x7a\xe6\xb6\x04",
.klen = 32,
.ptext = zeroed_string,
.ctext = "\x4f\x6a\x20\x38\x28\x68\x97\xb9"
"\xc9\x87\x01\x36\x55\x33\x17\xfa",
.len = 16,
}, { /* Generated from TF test vectors */
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
crypto: testmgr - expand serpent test vectors AVX2 implementation of serpent cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:57 -06:00
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xC3\x70\x22\x32\xF5\x80\xCB\x54"
"\xFC\x30\xE0\xF6\xEB\x39\x57\xA6"
"\xB6\xB9\xC5\xA4\x91\x55\x14\x97"
"\xC1\x20\xFF\x6C\x5C\xF0\x67\xEA"
"\x2F\xED\xD8\xC9\xFB\x38\x3F\xFE"
"\x93\xBE\xDC\x00\xD3\x7F\xAD\x4C"
"\x5A\x08\x92\xD1\x47\x0C\xFA\x6C"
"\xD0\x6A\x99\x10\x72\xF8\x47\x62"
"\x81\x42\xF8\xD8\xF5\xBB\x94\x08"
"\xAA\x97\xA2\x8B\x69\xB3\xD2\x7E"
"\xBC\xB5\x00\x0C\xE5\x44\x4B\x58"
"\xE8\x63\xDC\xB3\xC4\xE5\x23\x12"
"\x5A\x72\x85\x47\x8B\xEC\x9F\x26"
"\x84\xB6\xED\x10\x33\x63\x9B\x5F"
"\x4D\x53\xEE\x94\x45\x8B\x60\x58"
"\x86\x20\xF9\x1E\x82\x08\x3E\x58"
"\x60\x1B\x34\x19\x02\xBE\x4E\x09"
"\xBB\x7C\x15\xCC\x60\x27\x55\x7A"
"\x12\xB8\xD8\x08\x89\x3C\xA6\xF3"
"\xF1\xDD\xA7\x07\xA3\x12\x85\x28"
"\xE9\x57\xAC\x80\x0C\x5C\x0F\x3A"
"\x5D\xC2\x91\xC7\x90\xE4\x8C\x43"
"\x92\xE4\x7C\x26\x69\x4D\x83\x68"
"\x14\x96\x42\x47\xBD\xA9\xE4\x8A"
"\x33\x19\xEB\x54\x8E\x0D\x4B\x6E"
"\x91\x51\xB5\x36\x08\xDE\x1C\x06"
"\x03\xBD\xDE\x81\x26\xF7\x99\xC2"
"\xBA\xF7\x6D\x87\x0D\xE4\xA6\xCF"
"\xC1\xF5\x27\x05\xB8\x02\x57\x72"
"\xE6\x42\x13\x0B\xC6\x47\x05\x74"
"\x24\x15\xF7\x0D\xC2\x23\x9D\xB9"
"\x3C\x77\x18\x93\xBA\xB4\xFC\x8C"
"\x98\x82\x67\x67\xB4\xD7\xD3\x43"
"\x23\x08\x02\xB7\x9B\x99\x05\xFB"
"\xD3\xB5\x00\x0A\xA9\x9D\x66\xD6"
"\x2E\x49\x58\xD0\xA8\x57\x29\x7F"
"\x0A\x0E\x7D\xFC\x92\x83\xCC\x67"
"\xA2\xB1\x70\x3A\x8F\x87\x4A\x8D"
"\x17\xE2\x58\x2B\x88\x0D\x68\x62"
"\xBF\x35\xD1\x6F\xC0\xF0\x18\x62"
"\xB2\xC7\x2D\x58\xC7\x16\xDE\x08"
"\xEB\x84\x1D\x25\xA7\x38\x94\x06"
"\x93\x9D\xF8\xFE\x88\x71\xE7\x84"
"\x2C\xA0\x38\xA3\x1D\x48\xCF\x29"
"\x0B\xBC\xD8\x50\x99\x1A\x26\xFB"
"\x8E\x75\x3D\x73\xEB\x6A\xED\x29"
"\xE0\x8E\xED\xFC\xFE\x6F\xF6\xBA"
"\x41\xE2\x10\x4C\x01\x8B\x69\x2B"
"\x25\x3F\x4D\x70\x7B\x92\xD6\x3B"
"\xAC\xF9\x77\x18\xD9\x6A\x30\xA6"
"\x2E\xFA\x30\xFF\xC8\xD5\x1D\x06"
"\x59\x28\x1D\x86\x43\x04\x5D\x3B"
"\x99\x4C\x04\x5A\x21\x17\x8B\x76"
"\x8F\x72\xCB\xA1\x9C\x29\x4C\xC3"
"\x65\xA2\x58\x2A\xC5\x66\x24\xBF"
"\xBA\xE6\x0C\xDD\x34\x24\x74\xC8"
"\x84\x0A\x66\x2C\xBE\x8F\x32\xA9"
"\xE7\xE4\xA1\xD7\xDA\xAB\x23\x1E"
"\xEB\xEE\x6C\x94\x6F\x9C\x2E\xD1"
"\x49\x2C\xF3\xD4\x90\xCC\x93\x4C"
"\x84\x52\x6D\x68\xDE\xC6\x64\xB2"
"\x11\x74\x93\x57\xB4\x7E\xC6\x00",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cast6_cbc_tv_template[] = {
{ /* Generated from TF test vectors */
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
crypto: testmgr - expand serpent test vectors AVX2 implementation of serpent cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:57 -06:00
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xDF\x77\x68\x96\xC7\xBA\xF8\xE2"
"\x0E\x24\x99\x1A\xAA\xF3\xC6\x9F"
"\xA0\x73\xB3\x70\xC3\x68\x64\x70"
"\xAD\x33\x02\xFB\x88\x74\xAA\x78"
"\xC7\x47\x1A\x18\x61\x2D\xAC\x9F"
"\x7E\x6F\xDF\x05\x13\x76\xA6\x72"
"\xB7\x13\x09\x0F\x7D\x38\xDF\x25"
"\x4E\xFD\x50\x45\xFA\x35\x6A\xC0"
"\x57\x95\xE1\x21\x26\x10\x9A\x21"
"\xA1\x8A\x51\x05\xD1\xB1\x78\x35"
"\x98\xF5\xAE\xC0\xC1\x8B\x94\xFF"
"\xD0\x69\x3F\x42\xC2\x01\xA7\x9B"
"\x23\x16\x47\x72\x81\x13\x3A\x72"
"\xEC\xD9\x40\x88\x00\x9C\xB0\xA8"
"\x9C\xAC\xCE\x11\x73\x7B\x63\x3E"
"\xA3\x63\x98\x7D\x35\xE4\xD9\x83"
"\xE2\xD0\x52\x87\x0C\x1F\xB0\xB3"
"\x41\x1A\x93\x8D\x76\x31\x9F\xF2"
"\xFE\x09\xA3\x8F\x22\x6A\x3B\xB9"
"\x6C\x9E\xE4\xA1\xA0\xC4\xE7\xA1"
"\x21\x9C\x1A\xCA\x65\xDE\x44\x03"
"\x99\xF2\xD2\x39\xE3\x3F\x0F\x37"
"\x53\x50\x23\xA4\x81\x6E\xDA\xFB"
"\xF8\x7B\x01\xD7\xB2\x32\x9C\xB8"
"\xB1\x0E\x99\x17\xB5\x38\xF9\xD7"
"\x86\x2D\x6E\x94\x5C\x99\x9D\xB3"
"\xD3\x63\x4B\x2A\x7D\x44\x6A\xB2"
"\xC1\x03\xE6\x5A\x37\xD8\x64\x18"
"\xAA\x32\xCE\x29\xED\xC0\xA2\xCB"
"\x8D\xAF\xCD\xBE\x8F\xB6\xEC\xB4"
"\x89\x05\x81\x6E\x71\x4F\xC3\x28"
"\x10\xC1\x62\xC4\x41\xE9\xD2\x39"
"\xF3\x22\x39\x12\x2C\xC2\x95\x2D"
"\xBF\x93\x58\x4B\x04\xD1\x8D\x57"
"\xAE\xEB\x60\x03\x56\x35\xAD\x5A"
"\xE9\xC3\xFF\x4E\x31\xE1\x37\xF8"
"\x7D\xEE\x65\x8A\xB6\x88\x1A\x3E"
"\x07\x09\x82\xBA\xF0\x80\x8A\xD0"
"\xA0\x3F\x6A\xE9\x24\x87\x19\x65"
"\x73\x3F\x12\x91\x47\x54\xBA\x39"
"\x30\x5B\x1E\xE5\xC2\xF9\x3F\xEF"
"\xD6\x75\xF9\xB8\x7C\x8B\x05\x76"
"\xEE\xB7\x08\x25\x4B\xB6\x7B\x47"
"\x72\xC0\x4C\xD4\xDA\xE0\x75\xF1"
"\x7C\xE8\x94\x9E\x16\x6E\xB8\x12"
"\xA1\xC1\x6E\x3B\x1C\x59\x41\x2D"
"\x23\xFA\x7D\x77\xB8\x46\x75\xFE"
"\x4F\x10\xD3\x09\x60\xA1\x36\x96"
"\x5B\xC2\xDC\x6E\x84\x7D\x9B\x14"
"\x80\x21\x83\x58\x3C\x76\xFD\x28"
"\x1D\xF9\x93\x13\xD7\x0E\x62\x14"
"\x5A\xC5\x4E\x08\xA5\x56\xA4\x3C"
"\x68\x93\x44\x70\xDF\xCF\x4A\x51"
"\x0B\x81\x29\x41\xE5\x62\x4D\x36"
"\xB3\xEA\x94\xA6\xB9\xDD\x3F\x09"
"\x62\x34\xA0\x6A\x7E\x7D\xF5\xF6"
"\x01\x91\xB4\x27\xDA\x59\xD6\x17"
"\x56\x4D\x82\x62\x37\xA3\x48\x01"
"\x99\x91\x77\xB2\x08\x6B\x2C\x37"
"\xC5\x5C\xAD\xB6\x07\xB6\x84\xF3"
"\x4D\x59\x7D\xC5\x28\x69\xFA\x92"
"\x22\x46\x89\x2D\x0F\x2B\x08\x24",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cast6_ctr_tv_template[] = {
{ /* Generated from TF test vectors */
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x3A",
.ctext = "\x26\x0A\xF1\xE2\x3F\x8A\xEF\xA3"
"\x53\x9A\x5E\x1B\x2A\x1A\xC6\x0A"
"\x57",
.len = 17,
}, { /* Generated from TF test vectors */
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
crypto: testmgr - expand serpent test vectors AVX2 implementation of serpent cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:57 -06:00
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x26\x0A\xF1\xE2\x3F\x8A\xEF\xA3"
"\x53\x9A\x5E\x1B\x2A\x1A\xC6\x0A"
"\x57\xA3\xEF\x47\x2A\xE8\x88\xA7"
"\x3C\xD0\xEC\xB9\x94\x50\x7D\x56"
"\xBC\xE1\xC1\xF5\xE1\xEE\x12\xF8"
"\x4F\x03\x82\x3A\x93\x6B\x4C\xD3"
"\xE3\xF3\xFA\xC2\x23\x55\x98\x20"
"\x49\x76\x9B\x6B\xC1\x23\xBF\xE5"
"\xD4\xC4\x2F\x61\xE1\x67\x2A\x30"
"\x6F\x29\xCA\x54\xF8\x1B\xA6\x7D"
"\x66\x45\xEE\xC8\x19\xBE\x50\xF0"
"\x5F\x65\xF8\x1E\x4D\x07\x87\xD9"
"\xD3\xD9\x1B\x09\x89\xFD\x42\xC5"
"\xDB\xEB\x86\xF1\x67\x04\x0F\x5C"
"\x81\xDF\x82\x12\xC7\x4C\x1B\x07"
"\xDE\xE6\xFA\x29\x86\xD1\xB0\xBA"
"\x3D\x6A\x69\x76\xEC\x0F\xB4\xE6"
"\xCD\xA7\xF8\xA8\xB8\xE0\x33\xF5"
"\x49\x61\x22\x52\x64\x8C\x46\x41"
"\x1F\x48\x5F\x4F\xA2\x89\x36\x17"
"\x20\xF8\x2F\x8F\x4B\xFA\xF2\xC0"
"\x1E\x18\xA2\xF8\xB7\x6D\x98\xE3"
"\x00\x14\x15\x59\xC1\x30\x64\xAF"
"\xA8\x01\x38\xAB\xD4\x8B\xEC\x7C"
"\x44\x9A\xC6\x2C\x2E\x2B\x2B\xF4"
"\x02\x37\xC4\x69\xEF\x36\xC1\xF3"
"\xA0\xFB\xFE\x29\xAD\x39\xCF\xD0"
"\x51\x73\xA3\x22\x42\x41\xAB\xD2"
"\x0F\x50\x14\xB9\x54\xD3\xD4\xFA"
"\xBF\xC9\xBB\xCE\xC4\x1D\x2D\xAF"
"\xC9\x3F\x07\x87\x42\x4B\x3A\x54"
"\x34\x8E\x37\xA3\x03\x6F\x65\x66"
"\xDB\x44\xC3\xE8\xD7\xDD\x7D\xDD"
"\x61\xB4\x2B\x80\xA3\x98\x13\xF5"
"\x5A\xD3\x34\x58\xC3\x6E\xF6\xB8"
"\x0A\xC6\x50\x01\x8E\xD5\x6C\x7D"
"\xFE\x16\xB6\xCF\xFC\x51\x40\xAE"
"\xB3\x15\xAC\x90\x6F\x0B\x28\x3A"
"\x60\x40\x38\x90\x20\x46\xC7\xB3"
"\x0B\x12\x6D\x3B\x15\x14\xF9\xF4"
"\x11\x41\x76\x6B\xB3\x60\x82\x3C"
"\x84\xFB\x08\x2E\x92\x25\xCB\x79"
"\x6F\x58\xC5\x94\x00\x00\x47\xB6"
"\x9E\xDC\x0F\x29\x70\x46\x20\x76"
"\x65\x75\x66\x5C\x00\x96\xB3\xE1"
"\x0B\xA7\x11\x8B\x2E\x61\x4E\x45"
"\x73\xFC\x91\xAB\x79\x41\x23\x14"
"\x13\xB6\x72\x6C\x46\xB3\x03\x11"
"\xE4\xF1\xEE\xC9\x7A\xCF\x96\x32"
"\xB6\xF0\x8B\x97\xB4\xCF\x82\xB7"
"\x15\x48\x44\x99\x09\xF6\xE0\xD7"
"\xBC\xF1\x5B\x91\x4F\x30\x22\xA2"
"\x45\xC4\x68\x55\xC2\xBE\xA7\xD2"
"\x12\x53\x35\x9C\xF9\xE7\x35\x5D"
"\x81\xE4\x86\x42\xC3\x58\xFB\xF0"
"\x38\x9B\x8E\x5A\xEF\x83\x33\x0F"
"\x00\x4E\x3F\x9F\xF5\x84\x62\xC4"
"\x19\x35\x88\x22\x45\x59\x0E\x8F"
"\xEC\x27\xDD\x4A\xA4\x1F\xBC\x41"
"\x9B\x66\x8D\x32\xBA\x81\x34\x87"
"\x0E\x74\x33\x30\x62\xB9\x89\xDF"
"\xF9\xC5\xDD\x27\xB3\x39\xCB\xCB",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add new serpent test vectors Add new serpent tests for serpent_sse2 x86_64/i586 8-way/4-way code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-17 15:02:53 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cast6_lrw_tv_template[] = {
{ /* Generated from TF test vectors */
crypto: testmgr - add lrw(serpent) test vectors Add test vectors for lrw(serpent). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:34 -06:00
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x05\x11\xb7\x18\xab\xc6\x2d\xac"
crypto: testmgr - add lrw(serpent) test vectors Add test vectors for lrw(serpent). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:34 -06:00
"\x70\x5d\xf6\x22\x94\xcd\xe5\x6c"
"\x17\x6b\xf6\x1c\xf0\xf3\x6e\xf8"
"\x50\x38\x1f\x71\x49\xb6\x57\xd6"
"\x8f\xcb\x8d\x6b\xe3\xa6\x29\x90"
"\xfe\x2a\x62\x82\xae\x6d\x8b\xf6"
"\xad\x1e\x9e\x20\x5f\x38\xbe\x04"
"\xda\x10\x8e\xed\xa2\xa4\x87\xab"
"\xda\x6b\xb4\x0c\x75\xba\xd3\x7c"
"\xc9\xac\x42\x31\x95\x7c\xc9\x04"
"\xeb\xd5\x6e\x32\x69\x8a\xdb\xa6"
"\x15\xd7\x3f\x4f\x2f\x66\x69\x03"
"\x9c\x1f\x54\x0f\xde\x1f\xf3\x65"
"\x4c\x96\x12\xed\x7c\x92\x03\x01"
"\x6f\xbc\x35\x93\xac\xf1\x27\xf1"
"\xb4\x96\x82\x5a\x5f\xb0\xa0\x50"
"\x89\xa4\x8e\x66\x44\x85\xcc\xfd"
"\x33\x14\x70\xe3\x96\xb2\xc3\xd3"
"\xbb\x54\x5a\x1a\xf9\x74\xa2\xc5"
"\x2d\x64\x75\xdd\xb4\x54\xe6\x74"
"\x8c\xd3\x9d\x9e\x86\xab\x51\x53"
"\xb7\x93\x3e\x6f\xd0\x4e\x2c\x40"
"\xf6\xa8\x2e\x3e\x9d\xf4\x66\xa5"
"\x76\x12\x73\x44\x1a\x56\xd7\x72"
"\x88\xcd\x21\x8c\x4c\x0f\xfe\xda"
"\x95\xe0\x3a\xa6\xa5\x84\x46\xcd"
"\xd5\x3e\x9d\x3a\xe2\x67\xe6\x60"
"\x1a\xe2\x70\x85\x58\xc2\x1b\x09"
"\xe1\xd7\x2c\xca\xad\xa8\x8f\xf9"
"\xac\xb3\x0e\xdb\xca\x2e\xe2\xb8"
"\x51\x71\xd9\x3c\x6c\xf1\x56\xf8"
"\xea\x9c\xf1\xfb\x0c\xe6\xb7\x10"
"\x1c\xf8\xa9\x7c\xe8\x53\x35\xc1"
"\x90\x3e\x76\x4a\x74\xa4\x21\x2c"
"\xf6\x2c\x4e\x0f\x94\x3a\x88\x2e"
"\x41\x09\x6a\x33\x7d\xf6\xdd\x3f"
"\x8d\x23\x31\x74\x84\xeb\x88\x6e"
"\xcc\xb9\xbc\x22\x83\x19\x07\x22"
"\xa5\x2d\xdf\xa5\xf3\x80\x85\x78"
"\x84\x39\x6a\x6d\x6a\x99\x4f\xa5"
"\x15\xfe\x46\xb0\xe4\x6c\xa5\x41"
"\x3c\xce\x8f\x42\x60\x71\xa7\x75"
"\x08\x40\x65\x8a\x82\xbf\xf5\x43"
"\x71\x96\xa9\x4d\x44\x8a\x20\xbe"
"\xfa\x4d\xbb\xc0\x7d\x31\x96\x65"
"\xe7\x75\xe5\x3e\xfd\x92\x3b\xc9"
"\x55\xbb\x16\x7e\xf7\xc2\x8c\xa4"
"\x40\x1d\xe5\xef\x0e\xdf\xe4\x9a"
"\x62\x73\x65\xfd\x46\x63\x25\x3d"
"\x2b\xaf\xe5\x64\xfe\xa5\x5c\xcf"
"\x24\xf3\xb4\xac\x64\xba\xdf\x4b"
"\xc6\x96\x7d\x81\x2d\x8d\x97\xf7"
"\xc5\x68\x77\x84\x32\x2b\xcc\x85"
"\x74\x96\xf0\x12\x77\x61\xb9\xeb"
"\x71\xaa\x82\xcb\x1c\xdb\x89\xc8"
"\xc6\xb5\xe3\x5c\x7d\x39\x07\x24"
"\xda\x39\x87\x45\xc0\x2b\xbb\x01"
"\xac\xbc\x2a\x5c\x7f\xfc\xe8\xce"
"\x6d\x9c\x6f\xed\xd3\xc1\xa1\xd6"
"\xc5\x55\xa9\x66\x2f\xe1\xc8\x32"
"\xa6\x5d\xa4\x3a\x98\x73\xe8\x45"
"\xa4\xc7\xa8\xb4\xf6\x13\x03\xf6"
"\xe9\x2e\xc4\x29\x0f\x84\xdb\xc4"
"\x21\xc4\xc2\x75\x67\x89\x37\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x55\x25\x09\x8B\xB5\xD5\xF8\xBF"
"\x37\x4A\xFE\x3C\x47\xD8\xE6\xEB"
"\xCA\xA4\x9B\xB0\xAB\x6D\x64\xCA"
"\x58\xB6\x73\xF0\xD7\x52\x34\xEF"
"\xFB\x3E\x96\x81\xB7\x71\x34\xA4"
"\x55\x20\xBE\x39\x5A\x2B\xF9\xD1"
"\x65\x0B\xDA\xD3\x7E\xB3\xA6\xF7"
"\x2E\x0B\x5A\x52\xDB\x39\x8C\x9B"
"\x61\x17\x5F\xAF\xB6\x5A\xC8\x08"
"\xA7\xB7\x2A\x11\x7C\x97\x38\x9D"
"\x59\x0E\x66\x59\x5E\xD8\x8B\xCE"
"\x70\xE0\xC3\x42\xB0\x8C\x0F\xBA"
"\xB2\x0D\x81\xB6\xBE\x61\x1C\x2D"
"\x7E\xEA\x91\x25\xAC\xEC\xF8\x28"
"\x80\x1D\xF0\x30\xBA\x62\x77\x7D"
"\xDB\x15\x69\xDF\xFA\x2A\x81\x64"
"\x95\x5B\xA4\x7F\x3E\x4F\xE3\x30"
"\xB0\x5C\xC2\x05\xF8\xF0\x29\xE7"
"\x0A\xA0\x66\xB2\x5D\x0F\x39\x2B"
"\xB4\xB3\x00\xA9\xD0\xAB\x63\x61"
"\x5E\xDB\xFC\x11\x74\x25\x96\x65"
"\xE8\xE2\x34\x57\x77\x15\x5E\x70"
"\xFF\x10\x90\xC3\x64\xF0\x11\x0A"
"\x63\x3A\xD3\x55\x92\x15\x4B\x0C"
"\xC7\x08\x89\x17\x3B\x99\xAD\x63"
"\xE7\x06\xDF\x52\xBC\x15\x64\x45"
"\x9D\x7A\xFB\x69\xBC\x2D\x6E\xA9"
"\x35\xD9\xD8\xF5\x0C\xC4\xA2\x23"
"\x9C\x18\x8B\xA8\x8C\xFE\xF8\x0E"
"\xBD\xAB\x60\x1A\x51\x17\x54\x27"
"\xB6\xE8\xBE\x0F\xA9\xA5\x82\x19"
"\x2F\x6F\x20\xA7\x47\xED\x74\x6C"
"\x4E\xC1\xF8\x8C\x14\xF3\xBB\x1F"
"\xED\x4D\x8F\x7C\x37\xEF\x19\xA1"
"\x07\x16\xDE\x76\xCC\x5E\x94\x02"
"\xFB\xBF\xE4\x81\x50\xCE\xFC\x0F"
"\x9E\xCF\x3D\xF6\x67\x00\xBF\xA7"
"\x6E\x21\x58\x36\x06\xDE\xB3\xD4"
"\xA2\xFA\xD8\x4E\xE0\xB9\x7F\x23"
"\x51\x21\x2B\x32\x68\xAA\xF8\xA8"
"\x93\x08\xB5\x6D\xE6\x43\x2C\xB7"
"\x31\xB2\x0F\xD0\xA2\x51\xC0\x25"
"\x30\xC7\x10\x3F\x97\x27\x01\x8E"
"\xFA\xD8\x4F\x78\xD8\x2E\x1D\xEB"
"\xA1\x37\x52\x0F\x7B\x5E\x87\xA8"
"\x22\xE2\xE6\x92\xA7\x5F\x11\x32"
"\xCC\x93\x34\xFC\xD1\x7E\xAE\x54"
"\xBC\x6A\x1B\x91\xD1\x2E\x21\xEC"
"\x5D\xF1\xC4\xF1\x55\x20\xBF\xE5"
"\x96\x3D\x69\x91\x20\x4E\xF2\x61"
"\xDA\x77\xFE\xEE\xC3\x74\x57\x2A"
"\x78\x39\xB0\xE0\xCF\x12\x56\xD6"
"\x05\xDC\xF9\x19\x66\x44\x1D\xF9"
"\x82\x37\xD4\xC2\x60\xB6\x31\xDF"
"\x0C\xAF\xBC\x8B\x55\x9A\xC8\x2D"
"\xAB\xA7\x88\x7B\x41\xE8\x29\xC9"
"\x9B\x8D\xA7\x00\x86\x25\xB6\x14"
"\xF5\x13\x73\xD7\x4B\x6B\x83\xF3"
"\xAF\x96\x00\xE4\xB7\x3C\x65\xA6"
"\x15\xB7\x94\x7D\x4E\x70\x4C\x75"
"\xF3\xB4\x02\xA9\x17\x1C\x7A\x0A"
"\xC0\xD5\x33\x11\x56\xDE\xDC\xF5"
"\x8D\xD9\xCD\x3B\x22\x67\x18\xC7"
"\xC4\xF5\x99\x61\xBC\xBB\x5B\x46",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add lrw(serpent) test vectors Add test vectors for lrw(serpent). These are generated from lrw(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:32:34 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cast6_xts_tv_template[] = {
{ /* Generated from TF test vectors */
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xDE\x6F\x22\xA5\xE8\x39\xE8\x78"
"\x88\x5A\x4F\x8D\x82\x76\x52\x6D"
"\xB2\x41\x16\xF4\x2B\xA6\xEB\xF6"
"\xE2\xC5\x62\x8D\x61\xA1\x01\xED"
"\xD9\x38\x01\xC1\x43\x63\x4E\x88"
"\xC9\x4B\x5A\x88\x80\xB7\x5C\x71"
"\x47\xEE\x11\xD8\xB7\x2D\x5D\x13"
"\x1A\xB1\x68\x5B\x61\xA7\xA9\x81"
"\x8B\x83\xA1\x6A\xAA\x36\xD6\xB6"
"\x60\x54\x09\x32\xFE\x6A\x76\x2E"
"\x28\xFF\xD5\xD6\xDD\x1D\x45\x7D"
"\xF0\x8B\xF3\x32\x4E\x6C\x12\xCB"
"\xB8\x25\x70\xF8\x40\xBC\x90\x1B"
"\x11\xC3\x59\xAF\xF0\x2F\x92\xDD"
"\xD3\x3B\xCF\x60\xA1\x78\x94\x57"
"\xAF\x76\xC1\x67\xA6\x3C\xCD\x98"
"\xB1\xF7\x27\xB9\xA3\xBD\x10\xEA"
"\xCD\x8B\xC2\xF2\x14\xF2\xB2\x67"
"\x05\xDD\x1D\x58\x6E\x2F\x95\x08"
"\x3A\xF8\x78\x76\x82\x56\xA7\xEC"
"\x51\x4B\x85\x77\xC2\x4C\x4A\x34"
"\x71\x38\x17\x91\x44\xE8\xFC\x65"
"\x99\x0D\x52\x91\xEE\xF8\xEF\x27"
"\x2A\x9E\x6E\x78\xC4\x26\x87\xF4"
"\x8A\xF0\x2D\x04\xE8\x14\x92\x5D"
"\x59\x22\x9B\x29\x5C\x18\xF0\xC3"
"\x47\xF3\x76\xD8\xE4\xF3\x1B\xD1"
"\x70\xA3\x0D\xB5\x70\x02\x1D\xA3"
"\x91\x3B\x49\x73\x18\xAB\xD4\xC9"
"\xC3\x1E\xEF\x1F\xFE\xD5\x59\x8A"
"\xD7\xF6\xC9\x71\x67\x79\xD7\x0E"
"\xBE\x1F\x8E\xEC\x55\x7E\x4F\x24"
"\xE6\x87\xEA\xFE\x96\x25\x67\x8E"
"\x93\x03\xFA\xFF\xCE\xAF\xB2\x3C"
"\x6F\xEB\x57\xFB\xD3\x28\x87\xA9"
"\xCE\xC2\xF5\x9C\xC6\x67\xB5\x97"
"\x49\xF7\x04\xCB\xEF\x84\x98\x33"
"\xAF\x38\xD3\x04\x1C\x24\x71\x38"
"\xC7\x71\xDD\x43\x0D\x12\x4A\x18"
"\xBA\xC4\xAF\xBA\xB2\x5B\xEB\x95"
"\x02\x43\x5D\xCE\x19\xCC\xCD\x66"
"\x91\x0B\x8C\x7F\x51\xC4\xBF\x3C"
"\x8B\xF1\xCC\xAA\x29\xD7\x87\xCB"
"\x3E\xC5\xF3\xC9\x75\xE8\xA3\x5B"
"\x30\x45\xA9\xB7\xAF\x80\x64\x6F"
"\x75\x4A\xA7\xC0\x6D\x19\x6B\xDE"
"\x17\xDE\x6D\xEA\x87\x9F\x95\xAE"
"\xF5\x3C\xEE\x54\xB8\x27\x84\xF8"
"\x97\xA3\xE1\x6F\x38\x24\x34\x88"
"\xCE\xBD\x32\x52\xE0\x00\x6C\x94"
"\xC9\xD7\x5D\x37\x81\x33\x2E\x7F"
"\x4F\x7E\x2E\x0D\x94\xBD\xEA\x59"
"\x34\x39\xA8\x35\x12\xB7\xBC\xAC"
"\xEA\x52\x9C\x78\x02\x6D\x92\x36"
"\xFB\x59\x2B\xA4\xEA\x7B\x1B\x83"
"\xE1\x4D\x5E\x2A\x7E\x92\xB1\x64"
"\xDE\xE0\x27\x4B\x0A\x6F\x4C\xE3"
"\xB0\xEB\x31\xE4\x69\x95\xAB\x35"
"\x8B\x2C\xF5\x6B\x7F\xF1\xA2\x82"
"\xF8\xD9\x47\x82\xA9\x82\x03\x91"
"\x69\x1F\xBE\x4C\xE7\xC7\x34\x2F"
"\x45\x72\x80\x17\x81\xBD\x9D\x62"
"\xA1\xAC\xE8\xCF\xC6\x74\xCF\xDC"
"\x22\x60\x4E\xE8\xA4\x5D\x85\xB9",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* AES test vectors.
*/
static const struct cipher_testvec aes_tv_template[] = {
{ /* From FIPS-197 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.klen = 16,
.ptext = "\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff",
.ctext = "\x69\xc4\xe0\xd8\x6a\x7b\x04\x30"
"\xd8\xcd\xb7\x80\x70\xb4\xc5\x5a",
.len = 16,
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x10\x11\x12\x13\x14\x15\x16\x17",
.klen = 24,
.ptext = "\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff",
.ctext = "\xdd\xa9\x7c\xa4\x86\x4c\xdf\xe0"
"\x6e\xaf\x70\xa0\xec\x0d\x71\x91",
.len = 16,
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.klen = 32,
.ptext = "\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff",
.ctext = "\x8e\xa2\xb7\xca\x51\x67\x45\xbf"
"\xea\xfc\x49\x90\x4b\x49\x60\x89",
.len = 16,
}, { /* Generated with Crypto++ */
.key = "\xA6\xC9\x83\xA6\xC9\xEC\x0F\x32"
"\x55\x0F\x32\x55\x78\x9B\xBE\x78"
"\x9B\xBE\xE1\x04\x27\xE1\x04\x27"
"\x4A\x6D\x90\x4A\x6D\x90\xB3\xD6",
.klen = 32,
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB"
"\x54\xE0\x49\xB2\x1B\xA7\x10\x79"
"\x05\x6E\xD7\x40\xCC\x35\x9E\x07"
"\x93\xFC\x65\xF1\x5A\xC3\x2C\xB8"
"\x21\x8A\x16\x7F\xE8\x51\xDD\x46"
"\xAF\x18\xA4\x0D\x76\x02\x6B\xD4"
"\x3D\xC9\x32\x9B\x04\x90\xF9\x62"
"\xEE\x57\xC0\x29\xB5\x1E\x87\x13"
"\x7C\xE5\x4E\xDA\x43\xAC\x15\xA1"
"\x0A\x73\xFF\x68\xD1\x3A\xC6\x2F"
"\x98\x01\x8D\xF6\x5F\xEB\x54\xBD"
"\x26\xB2\x1B\x84\x10\x79\xE2\x4B"
"\xD7\x40\xA9\x12\x9E\x07\x70\xFC"
"\x65\xCE\x37\xC3\x2C\x95\x21\x8A"
"\xF3\x5C\xE8\x51\xBA\x23\xAF\x18"
"\x81\x0D\x76\xDF\x48\xD4\x3D\xA6"
"\x0F\x9B\x04\x6D\xF9\x62\xCB\x34"
"\xC0\x29\x92\x1E\x87\xF0\x59\xE5"
"\x4E\xB7\x20\xAC\x15\x7E\x0A\x73"
"\xDC\x45\xD1\x3A\xA3\x0C\x98\x01"
"\x6A\xF6\x5F\xC8\x31\xBD\x26\x8F"
"\x1B\x84\xED\x56\xE2\x4B\xB4\x1D"
"\xA9\x12\x7B\x07\x70\xD9\x42\xCE"
"\x37\xA0\x09\x95\xFE\x67\xF3\x5C"
"\xC5\x2E\xBA\x23\x8C\x18\x81\xEA"
"\x53\xDF\x48\xB1\x1A\xA6\x0F\x78"
"\x04\x6D\xD6\x3F\xCB\x34\x9D\x06"
"\x92\xFB\x64\xF0\x59\xC2\x2B\xB7"
"\x20\x89\x15\x7E\xE7\x50\xDC\x45"
"\xAE\x17\xA3\x0C\x75\x01\x6A\xD3"
"\x3C\xC8\x31\x9A\x03\x8F\xF8\x61"
"\xED\x56\xBF\x28\xB4\x1D\x86\x12",
.ctext = "\x71\x73\xF7\xDB\x24\x93\x21\x6D"
"\x61\x1E\xBB\x63\x42\x79\xDB\x64"
"\x6F\x82\xC0\xCA\xA3\x9B\xFA\x0B"
"\xD9\x08\xC7\x4A\x90\xAE\x8F\x5F"
"\x5E\x06\xF0\x5F\x31\x51\x18\x37"
"\x45\xD7\xCA\x3A\xFD\x6C\x3F\xE1"
"\xDD\x8D\x22\x65\x2B\x00\x50\xCE"
"\xBA\x28\x67\xD7\xCE\x0E\x0D\xEA"
"\x78\x69\x7F\xAE\x8F\x8B\x69\x37"
"\x75\xE0\xDC\x96\xE0\xB7\xF4\x09"
"\xCB\x6D\xA2\xFB\xDA\xAF\x09\xF8"
"\x81\x82\x27\xFA\x45\x9C\x29\xA4"
"\x22\x8B\x78\x69\x5B\x46\xF9\x39"
"\x1B\xCC\xF9\x1D\x09\xEB\xBC\x5C"
"\x41\x72\x51\x97\x1D\x07\x49\xA0"
"\x1B\x8E\x65\x4B\xB2\x6A\x12\x03"
"\x6A\x60\x95\xAC\xBD\xAC\x1A\x64"
"\xDE\x5A\xA5\xF0\x83\x2F\xCB\xCA"
"\x22\x74\xA6\x6C\x9B\x73\xCE\x3F"
"\xE1\x8B\x22\x17\x59\x0C\x47\x89"
"\x33\xA1\xD6\x47\x03\x19\x4F\xA8"
"\x67\x69\xF0\x5B\xF0\x20\xAD\x06"
"\x27\x81\x92\xD8\xC5\xBA\x98\x12"
"\xBE\x24\xB5\x2F\x75\x02\xC2\xAD"
"\x12\x2F\x07\x32\xEE\x39\xAF\x64"
"\x05\x8F\xB3\xD4\xEB\x1B\x46\x6E"
"\xD9\x21\xF9\xC4\xB7\xC9\x45\x68"
"\xB4\xA1\x74\x9F\x82\x47\xEB\xCC"
"\xBD\x0A\x14\x95\x0F\x8B\xA8\x2F"
"\x4B\x1B\xA7\xBF\x82\xA6\x43\x0C"
"\xB9\x39\x4A\xA8\x10\x6F\x50\x7B"
"\x25\xFB\x26\x81\xE0\x2F\xF0\x96"
"\x8D\x8B\xAC\x92\x0F\xF6\xED\x64"
"\x63\x29\x4C\x8E\x18\x13\xC5\xBF"
"\xFC\xA0\xD9\xBF\x7C\x3A\x0E\x29"
"\x6F\xD1\x6C\x6F\xA5\xDA\xBF\xB1"
"\x30\xEA\x44\x2D\xC3\x8F\x16\xE1"
"\x66\xFA\xA3\x21\x3E\xFC\x13\xCA"
"\xF0\xF6\xF0\x59\xBD\x8F\x38\x50"
"\x31\xCB\x69\x3F\x96\x15\xD6\xF5"
"\xAE\xFF\xF6\xAA\x41\x85\x4C\x10"
"\x58\xE3\xF9\x44\xE6\x28\xDA\x9A"
"\xDC\x6A\x80\x34\x73\x97\x1B\xC5"
"\xCA\x26\x16\x77\x0E\x60\xAB\x89"
"\x0F\x04\x27\xBD\xCE\x3E\x71\xB4"
"\xA0\xD7\x22\x7E\xDB\xEB\x24\x70"
"\x42\x71\x51\x78\x70\xB3\xE0\x3D"
"\x84\x8E\x8D\x7B\xD0\x6D\xEA\x92"
"\x11\x08\x42\x4F\xE5\xAD\x26\x92"
"\xD2\x00\xAE\xA8\xE3\x4B\x37\x47"
"\x22\xC1\x95\xC1\x63\x7F\xCB\x03"
"\xF3\xE3\xD7\x9D\x60\xC7\xBC\xEA"
"\x35\xA2\xFD\x45\x52\x39\x13\x6F"
"\xC1\x53\xF3\x53\xDF\x33\x84\xD7"
"\xD2\xC8\x37\xB0\x75\xE3\x41\x46"
"\xB3\xC7\x83\x2E\x8A\xBB\xA4\xE5"
"\x7F\x3C\xFD\x8B\xEB\xEA\x63\xBD"
"\xB7\x46\xE7\xBF\x09\x9C\x0D\x0F"
"\x40\x86\x7F\x51\xE1\x11\x9C\xCB"
"\x88\xE6\x68\x47\xE3\x2B\xC5\xFF"
"\x09\x79\xA0\x43\x5C\x0D\x08\x58"
"\x17\xBB\xC0\x6B\x62\x3F\x56\xE9",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
};
static const struct cipher_testvec aes_cbc_tv_template[] = {
{ /* From RFC 3602 */
.key = "\x06\xa9\x21\x40\x36\xb8\xa1\x5b"
"\x51\x2e\x03\xd5\x34\x12\x00\x06",
.klen = 16,
.iv = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.ptext = "Single block msg",
.ctext = "\xe3\x53\x77\x9c\x10\x79\xae\xb8"
"\x27\x08\x94\x2d\xbe\x77\x18\x1a",
.len = 16,
.also_non_np = 1,
.np = 8,
.tap = { 3, 2, 3, 2, 3, 1, 1, 1 },
crypto: testmgr - add xts(serpent) test vectors Add test vectors for xts(serpent). These are generated from xts(aes) test vectors. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-10-18 04:33:17 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc2\x86\x69\x6d\x88\x7c\x9a\xa0"
"\x61\x1b\xbb\x3e\x20\x25\xa4\x5a",
.klen = 16,
.iv = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xd2\x96\xcd\x94\xc2\xcc\xcf\x8a"
"\x3a\x86\x30\x28\xb5\xe1\xdc\x0a"
"\x75\x86\x60\x2d\x25\x3c\xff\xf9"
"\x1b\x82\x66\xbe\xa6\xd6\x1a\xb1",
.len = 32,
}, { /* From NIST SP800-38A */
.key = "\x8e\x73\xb0\xf7\xda\x0e\x64\x52"
"\xc8\x10\xf3\x2b\x80\x90\x79\xe5"
"\x62\xf8\xea\xd2\x52\x2c\x6b\x7b",
.klen = 24,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\x4f\x02\x1d\xb2\x43\xbc\x63\x3d"
"\x71\x78\x18\x3a\x9f\xa0\x71\xe8"
"\xb4\xd9\xad\xa9\xad\x7d\xed\xf4"
"\xe5\xe7\x38\x76\x3f\x69\x14\x5a"
"\x57\x1b\x24\x20\x12\xfb\x7a\xe0"
"\x7f\xa9\xba\xac\x3d\xf1\x02\xe0"
"\x08\xb0\xe2\x79\x88\x59\x88\x81"
"\xd9\x20\xa9\xe6\x4f\x56\x15\xcd",
.len = 64,
}, {
.key = "\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\xf5\x8c\x4c\x04\xd6\xe5\xf1\xba"
"\x77\x9e\xab\xfb\x5f\x7b\xfb\xd6"
"\x9c\xfc\x4e\x96\x7e\xdb\x80\x8d"
"\x67\x9f\x77\x7b\xc6\x70\x2c\x7d"
"\x39\xf2\x33\x69\xa9\xd9\xba\xcf"
"\xa5\x30\xe2\x63\x04\x23\x14\x61"
"\xb2\xeb\x05\xe2\xc3\x9b\xe9\xfc"
"\xda\x6c\x19\x07\x8c\x6a\x9d\x1b",
.len = 64,
}, { /* Generated with Crypto++ */
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55"
"\x0F\x32\x55\x78\x9B\xBE\x78\x9B"
"\xBE\xE1\x04\x27\xE1\x04\x27\x4A"
"\x6D\x90\x4A\x6D\x90\xB3\xD6\xF9",
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE7\x82\x1D\xB8\x53\x11\xAC\x47"
"\xE2\x7D\x18\xD6\x71\x0C\xA7\x42",
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB"
"\x54\xE0\x49\xB2\x1B\xA7\x10\x79"
"\x05\x6E\xD7\x40\xCC\x35\x9E\x07"
"\x93\xFC\x65\xF1\x5A\xC3\x2C\xB8"
"\x21\x8A\x16\x7F\xE8\x51\xDD\x46"
"\xAF\x18\xA4\x0D\x76\x02\x6B\xD4"
"\x3D\xC9\x32\x9B\x04\x90\xF9\x62"
"\xEE\x57\xC0\x29\xB5\x1E\x87\x13"
"\x7C\xE5\x4E\xDA\x43\xAC\x15\xA1"
"\x0A\x73\xFF\x68\xD1\x3A\xC6\x2F"
"\x98\x01\x8D\xF6\x5F\xEB\x54\xBD"
"\x26\xB2\x1B\x84\x10\x79\xE2\x4B"
"\xD7\x40\xA9\x12\x9E\x07\x70\xFC"
"\x65\xCE\x37\xC3\x2C\x95\x21\x8A"
"\xF3\x5C\xE8\x51\xBA\x23\xAF\x18"
"\x81\x0D\x76\xDF\x48\xD4\x3D\xA6"
"\x0F\x9B\x04\x6D\xF9\x62\xCB\x34"
"\xC0\x29\x92\x1E\x87\xF0\x59\xE5"
"\x4E\xB7\x20\xAC\x15\x7E\x0A\x73"
"\xDC\x45\xD1\x3A\xA3\x0C\x98\x01"
"\x6A\xF6\x5F\xC8\x31\xBD\x26\x8F"
"\x1B\x84\xED\x56\xE2\x4B\xB4\x1D"
"\xA9\x12\x7B\x07\x70\xD9\x42\xCE"
"\x37\xA0\x09\x95\xFE\x67\xF3\x5C"
"\xC5\x2E\xBA\x23\x8C\x18\x81\xEA"
"\x53\xDF\x48\xB1\x1A\xA6\x0F\x78"
"\x04\x6D\xD6\x3F\xCB\x34\x9D\x06"
"\x92\xFB\x64\xF0\x59\xC2\x2B\xB7"
"\x20\x89\x15\x7E\xE7\x50\xDC\x45"
"\xAE\x17\xA3\x0C\x75\x01\x6A\xD3"
"\x3C\xC8\x31\x9A\x03\x8F\xF8\x61"
"\xED\x56\xBF\x28\xB4\x1D\x86\x12",
.ctext = "\xEA\x65\x8A\x19\xB0\x66\xC1\x3F"
"\xCE\xF1\x97\x75\xC1\xFD\xB5\xAF"
"\x52\x65\xF7\xFF\xBC\xD8\x2D\x9F"
"\x2F\xB9\x26\x9B\x6F\x10\xB7\xB8"
"\x26\xA1\x02\x46\xA2\xAD\xC6\xC0"
"\x11\x15\xFF\x6D\x1E\x82\x04\xA6"
"\xB1\x74\xD1\x08\x13\xFD\x90\x7C"
"\xF5\xED\xD3\xDB\x5A\x0A\x0C\x2F"
"\x0A\x70\xF1\x88\x07\xCF\x21\x26"
"\x40\x40\x8A\xF5\x53\xF7\x24\x4F"
"\x83\x38\x43\x5F\x08\x99\xEB\xE3"
"\xDC\x02\x64\x67\x50\x6E\x15\xC3"
"\x01\x1A\xA0\x81\x13\x65\xA6\x73"
"\x71\xA6\x3B\x91\x83\x77\xBE\xFA"
"\xDB\x71\x73\xA6\xC1\xAE\x43\xC3"
"\x36\xCE\xD6\xEB\xF9\x30\x1C\x4F"
"\x80\x38\x5E\x9C\x6E\xAB\x98\x2F"
"\x53\xAF\xCF\xC8\x9A\xB8\x86\x43"
"\x3E\x86\xE7\xA1\xF4\x2F\x30\x40"
"\x03\xA8\x6C\x50\x42\x9F\x77\x59"
"\x89\xA0\xC5\xEC\x9A\xB8\xDD\x99"
"\x16\x24\x02\x07\x48\xAE\xF2\x31"
"\x34\x0E\xC3\x85\xFE\x1C\x95\x99"
"\x87\x58\x98\x8B\xE7\xC6\xC5\x70"
"\x73\x81\x07\x7C\x56\x2F\xD8\x1B"
"\xB7\xB9\x2B\xAB\xE3\x01\x87\x0F"
"\xD8\xBB\xC0\x0D\xAC\x2C\x2F\x98"
"\x3C\x0B\xA2\x99\x4A\x8C\xF7\x04"
"\xE0\xE0\xCF\xD1\x81\x5B\xFE\xF5"
"\x24\x04\xFD\xB8\xDF\x13\xD8\xCD"
"\xF1\xE3\x3D\x98\x50\x02\x77\x9E"
"\xBC\x22\xAB\xFA\xC2\x43\x1F\x66"
"\x20\x02\x23\xDA\xDF\xA0\x89\xF6"
"\xD8\xF3\x45\x24\x53\x6F\x16\x77"
"\x02\x3E\x7B\x36\x5F\xA0\x3B\x78"
"\x63\xA2\xBD\xB5\xA4\xCA\x1E\xD3"
"\x57\xBC\x0B\x9F\x43\x51\x28\x4F"
"\x07\x50\x6C\x68\x12\x07\xCF\xFA"
"\x6B\x72\x0B\xEB\xF8\x88\x90\x2C"
"\x7E\xF5\x91\xD1\x03\xD8\xD5\xBD"
"\x22\x39\x7B\x16\x03\x01\x69\xAF"
"\x3D\x38\x66\x28\x0C\xBE\x5B\xC5"
"\x03\xB4\x2F\x51\x8A\x56\x17\x2B"
"\x88\x42\x6D\x40\x68\x8F\xD0\x11"
"\x19\xF9\x1F\x43\x79\x95\x31\xFA"
"\x28\x7A\x3D\xF7\x66\xEB\xEF\xAC"
"\x06\xB2\x01\xAD\xDB\x68\xDB\xEC"
"\x8D\x53\x6E\x72\x68\xA3\xC7\x63"
"\x43\x2B\x78\xE0\x04\x29\x8F\x72"
"\xB2\x2C\xE6\x84\x03\x30\x6D\xCD"
"\x26\x92\x37\xE1\x2F\xBB\x8B\x9D"
"\xE4\x4C\xF6\x93\xBC\xD9\xAD\x44"
"\x52\x65\xC7\xB0\x0E\x3F\x0E\x61"
"\x56\x5D\x1C\x6D\xA7\x05\x2E\xBC"
"\x58\x08\x15\xAB\x12\xAB\x17\x4A"
"\x5E\x1C\xF2\xCD\xB8\xA2\xAE\xFB"
"\x9B\x2E\x0E\x85\x34\x80\x0E\x3F"
"\x4C\xB8\xDB\xCE\x1C\x90\xA1\x61"
"\x6C\x69\x09\x35\x9E\xD4\xF4\xAD"
"\xBC\x06\x41\xE3\x01\xB4\x4E\x0A"
"\xE0\x1F\x91\xF8\x82\x96\x2D\x65"
"\xA3\xAA\x13\xCC\x50\xFF\x7B\x02",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_md5_ecb_cipher_null_enc_tv_template[] = {
{ /* Input data from RFC 2410 Case 1 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 8 + 16 + 0,
.iv = "",
.input = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ilen = 8,
.result = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xaa\x42\xfe\x43\x8d\xea\xa3\x5a"
"\xb9\x3d\x9f\xb1\xa3\x8e\x9b\xae",
.rlen = 8 + 16,
}, { /* Input data from RFC 2410 Case 2 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 8 + 16 + 0,
.iv = "",
.input = "Network Security People Have A Strange Sense Of Humor",
.ilen = 53,
.result = "Network Security People Have A Strange Sense Of Humor"
"\x73\xa5\x3e\x1c\x08\x0e\x8a\x8a"
"\x8e\xb5\x5f\x90\x8e\xfe\x13\x23",
.rlen = 53 + 16,
},
};
static const struct aead_testvec hmac_md5_ecb_cipher_null_dec_tv_template[] = {
{
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 8 + 16 + 0,
.iv = "",
.input = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xaa\x42\xfe\x43\x8d\xea\xa3\x5a"
"\xb9\x3d\x9f\xb1\xa3\x8e\x9b\xae",
.ilen = 8 + 16,
.result = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.rlen = 8,
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 8 + 16 + 0,
.iv = "",
.input = "Network Security People Have A Strange Sense Of Humor"
"\x73\xa5\x3e\x1c\x08\x0e\x8a\x8a"
"\x8e\xb5\x5f\x90\x8e\xfe\x13\x23",
.ilen = 53 + 16,
.result = "Network Security People Have A Strange Sense Of Humor",
.rlen = 53,
},
};
static const struct aead_testvec hmac_sha1_aes_cbc_enc_tv_temp[] = {
{ /* RFC 3602 Case 1 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00"
"\x06\xa9\x21\x40\x36\xb8\xa1\x5b"
"\x51\x2e\x03\xd5\x34\x12\x00\x06",
.klen = 8 + 20 + 16,
.iv = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.assoc = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.alen = 16,
.input = "Single block msg",
.ilen = 16,
.result = "\xe3\x53\x77\x9c\x10\x79\xae\xb8"
"\x27\x08\x94\x2d\xbe\x77\x18\x1a"
"\x1b\x13\xcb\xaf\x89\x5e\xe1\x2c"
"\x13\xc5\x2e\xa3\xcc\xed\xdc\xb5"
"\x03\x71\xa2\x06",
.rlen = 16 + 20,
}, { /* RFC 3602 Case 2 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x30\x31\x32\x33"
"\xc2\x86\x69\x6d\x88\x7c\x9a\xa0"
"\x61\x1b\xbb\x3e\x20\x25\xa4\x5a",
.klen = 8 + 20 + 16,
.iv = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.assoc = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.alen = 16,
.input = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.ilen = 32,
.result = "\xd2\x96\xcd\x94\xc2\xcc\xcf\x8a"
"\x3a\x86\x30\x28\xb5\xe1\xdc\x0a"
"\x75\x86\x60\x2d\x25\x3c\xff\xf9"
"\x1b\x82\x66\xbe\xa6\xd6\x1a\xb1"
"\xad\x9b\x4c\x5c\x85\xe1\xda\xae"
"\xee\x81\x4e\xd7\xdb\x74\xcf\x58"
"\x65\x39\xf8\xde",
.rlen = 32 + 20,
}, { /* RFC 3602 Case 3 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\x6c\x3e\xa0\x47\x76\x30\xce\x21"
"\xa2\xce\x33\x4a\xa7\x46\xc2\xcd",
.klen = 8 + 20 + 16,
.iv = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.assoc = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.alen = 16,
.input = "This is a 48-byte message (exactly 3 AES blocks)",
.ilen = 48,
.result = "\xd0\xa0\x2b\x38\x36\x45\x17\x53"
"\xd4\x93\x66\x5d\x33\xf0\xe8\x86"
"\x2d\xea\x54\xcd\xb2\x93\xab\xc7"
"\x50\x69\x39\x27\x67\x72\xf8\xd5"
"\x02\x1c\x19\x21\x6b\xad\x52\x5c"
"\x85\x79\x69\x5d\x83\xba\x26\x84"
"\xc2\xec\x0c\xf8\x7f\x05\xba\xca"
"\xff\xee\x4c\xd0\x93\xe6\x36\x7f"
"\x8d\x62\xf2\x1e",
.rlen = 48 + 20,
}, { /* RFC 3602 Case 4 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\x56\xe4\x7a\x38\xc5\x59\x89\x74"
"\xbc\x46\x90\x3d\xba\x29\x03\x49",
.klen = 8 + 20 + 16,
.iv = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.assoc = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.alen = 16,
.input = "\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf",
.ilen = 64,
.result = "\xc3\x0e\x32\xff\xed\xc0\x77\x4e"
"\x6a\xff\x6a\xf0\x86\x9f\x71\xaa"
"\x0f\x3a\xf0\x7a\x9a\x31\xa9\xc6"
"\x84\xdb\x20\x7e\xb0\xef\x8e\x4e"
"\x35\x90\x7a\xa6\x32\xc3\xff\xdf"
"\x86\x8b\xb7\xb2\x9d\x3d\x46\xad"
"\x83\xce\x9f\x9a\x10\x2e\xe9\x9d"
"\x49\xa5\x3e\x87\xf4\xc3\xda\x55"
"\x1c\x45\x57\xa9\x56\xcb\xa9\x2d"
"\x18\xac\xf1\xc7\x5d\xd1\xcd\x0d"
"\x1d\xbe\xc6\xe9",
.rlen = 64 + 20,
}, { /* RFC 3602 Case 5 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\x90\xd3\x82\xb4\x10\xee\xba\x7a"
"\xd9\x38\xc4\x6c\xec\x1a\x82\xbf",
.klen = 8 + 20 + 16,
.iv = "\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.alen = 24,
.input = "\x08\x00\x0e\xbd\xa7\x0a\x00\x00"
"\x8e\x9c\x08\x3d\xb9\x5b\x07\x00"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0e\x01",
.ilen = 80,
.result = "\xf6\x63\xc2\x5d\x32\x5c\x18\xc6"
"\xa9\x45\x3e\x19\x4e\x12\x08\x49"
"\xa4\x87\x0b\x66\xcc\x6b\x99\x65"
"\x33\x00\x13\xb4\x89\x8d\xc8\x56"
"\xa4\x69\x9e\x52\x3a\x55\xdb\x08"
"\x0b\x59\xec\x3a\x8e\x4b\x7e\x52"
"\x77\x5b\x07\xd1\xdb\x34\xed\x9c"
"\x53\x8a\xb5\x0c\x55\x1b\x87\x4a"
"\xa2\x69\xad\xd0\x47\xad\x2d\x59"
"\x13\xac\x19\xb7\xcf\xba\xd4\xa6"
"\x58\xc6\x84\x75\xe4\xe9\x6b\x0c"
"\xe1\xc5\x0b\x73\x4d\x82\x55\xa8"
"\x85\xe1\x59\xf7",
.rlen = 80 + 20,
}, { /* NIST SP800-38A F.2.3 CBC-AES192.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\x8e\x73\xb0\xf7\xda\x0e\x64\x52"
"\xc8\x10\xf3\x2b\x80\x90\x79\xe5"
"\x62\xf8\xea\xd2\x52\x2c\x6b\x7b",
.klen = 8 + 20 + 24,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\x4f\x02\x1d\xb2\x43\xbc\x63\x3d"
"\x71\x78\x18\x3a\x9f\xa0\x71\xe8"
"\xb4\xd9\xad\xa9\xad\x7d\xed\xf4"
"\xe5\xe7\x38\x76\x3f\x69\x14\x5a"
"\x57\x1b\x24\x20\x12\xfb\x7a\xe0"
"\x7f\xa9\xba\xac\x3d\xf1\x02\xe0"
"\x08\xb0\xe2\x79\x88\x59\x88\x81"
"\xd9\x20\xa9\xe6\x4f\x56\x15\xcd"
"\x73\xe3\x19\x3f\x8b\xc9\xc6\xf4"
"\x5a\xf1\x5b\xa8\x98\x07\xc5\x36"
"\x47\x4c\xfc\x36",
.rlen = 64 + 20,
}, { /* NIST SP800-38A F.2.5 CBC-AES256.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x20" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.klen = 8 + 20 + 32,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\xf5\x8c\x4c\x04\xd6\xe5\xf1\xba"
"\x77\x9e\xab\xfb\x5f\x7b\xfb\xd6"
"\x9c\xfc\x4e\x96\x7e\xdb\x80\x8d"
"\x67\x9f\x77\x7b\xc6\x70\x2c\x7d"
"\x39\xf2\x33\x69\xa9\xd9\xba\xcf"
"\xa5\x30\xe2\x63\x04\x23\x14\x61"
"\xb2\xeb\x05\xe2\xc3\x9b\xe9\xfc"
"\xda\x6c\x19\x07\x8c\x6a\x9d\x1b"
"\xa3\xe8\x9b\x17\xe3\xf4\x7f\xde"
"\x1b\x9f\xc6\x81\x26\x43\x4a\x87"
"\x51\xee\xd6\x4e",
.rlen = 64 + 20,
},
};
static const struct aead_testvec hmac_sha1_ecb_cipher_null_enc_tv_temp[] = {
{ /* Input data from RFC 2410 Case 1 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 8 + 20 + 0,
.iv = "",
.input = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ilen = 8,
.result = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\x40\xc3\x0a\xa1\xc9\xa0\x28\xab"
"\x99\x5e\x19\x04\xd1\x72\xef\xb8"
"\x8c\x5e\xe4\x08",
.rlen = 8 + 20,
}, { /* Input data from RFC 2410 Case 2 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 8 + 20 + 0,
.iv = "",
.input = "Network Security People Have A Strange Sense Of Humor",
.ilen = 53,
.result = "Network Security People Have A Strange Sense Of Humor"
"\x75\x6f\x42\x1e\xf8\x50\x21\xd2"
"\x65\x47\xee\x8e\x1a\xef\x16\xf6"
"\x91\x56\xe4\xd6",
.rlen = 53 + 20,
},
};
static const struct aead_testvec hmac_sha1_ecb_cipher_null_dec_tv_temp[] = {
{
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 8 + 20 + 0,
.iv = "",
.input = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\x40\xc3\x0a\xa1\xc9\xa0\x28\xab"
"\x99\x5e\x19\x04\xd1\x72\xef\xb8"
"\x8c\x5e\xe4\x08",
.ilen = 8 + 20,
.result = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.rlen = 8,
}, {
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x00" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 8 + 20 + 0,
.iv = "",
.input = "Network Security People Have A Strange Sense Of Humor"
"\x75\x6f\x42\x1e\xf8\x50\x21\xd2"
"\x65\x47\xee\x8e\x1a\xef\x16\xf6"
"\x91\x56\xe4\xd6",
.ilen = 53 + 20,
.result = "Network Security People Have A Strange Sense Of Humor",
.rlen = 53,
},
};
static const struct aead_testvec hmac_sha256_aes_cbc_enc_tv_temp[] = {
{ /* RFC 3602 Case 1 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x06\xa9\x21\x40\x36\xb8\xa1\x5b"
"\x51\x2e\x03\xd5\x34\x12\x00\x06",
.klen = 8 + 32 + 16,
.iv = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.assoc = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.alen = 16,
.input = "Single block msg",
.ilen = 16,
.result = "\xe3\x53\x77\x9c\x10\x79\xae\xb8"
"\x27\x08\x94\x2d\xbe\x77\x18\x1a"
"\xcc\xde\x2d\x6a\xae\xf1\x0b\xcc"
"\x38\x06\x38\x51\xb4\xb8\xf3\x5b"
"\x5c\x34\xa6\xa3\x6e\x0b\x05\xe5"
"\x6a\x6d\x44\xaa\x26\xa8\x44\xa5",
.rlen = 16 + 32,
}, { /* RFC 3602 Case 2 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xc2\x86\x69\x6d\x88\x7c\x9a\xa0"
"\x61\x1b\xbb\x3e\x20\x25\xa4\x5a",
.klen = 8 + 32 + 16,
.iv = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.assoc = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.alen = 16,
.input = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.ilen = 32,
.result = "\xd2\x96\xcd\x94\xc2\xcc\xcf\x8a"
"\x3a\x86\x30\x28\xb5\xe1\xdc\x0a"
"\x75\x86\x60\x2d\x25\x3c\xff\xf9"
"\x1b\x82\x66\xbe\xa6\xd6\x1a\xb1"
"\xf5\x33\x53\xf3\x68\x85\x2a\x99"
"\x0e\x06\x58\x8f\xba\xf6\x06\xda"
"\x49\x69\x0d\x5b\xd4\x36\x06\x62"
"\x35\x5e\x54\x58\x53\x4d\xdf\xbf",
.rlen = 32 + 32,
}, { /* RFC 3602 Case 3 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x6c\x3e\xa0\x47\x76\x30\xce\x21"
"\xa2\xce\x33\x4a\xa7\x46\xc2\xcd",
.klen = 8 + 32 + 16,
.iv = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.assoc = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.alen = 16,
.input = "This is a 48-byte message (exactly 3 AES blocks)",
.ilen = 48,
.result = "\xd0\xa0\x2b\x38\x36\x45\x17\x53"
"\xd4\x93\x66\x5d\x33\xf0\xe8\x86"
"\x2d\xea\x54\xcd\xb2\x93\xab\xc7"
"\x50\x69\x39\x27\x67\x72\xf8\xd5"
"\x02\x1c\x19\x21\x6b\xad\x52\x5c"
"\x85\x79\x69\x5d\x83\xba\x26\x84"
"\x68\xb9\x3e\x90\x38\xa0\x88\x01"
"\xe7\xc6\xce\x10\x31\x2f\x9b\x1d"
"\x24\x78\xfb\xbe\x02\xe0\x4f\x40"
"\x10\xbd\xaa\xc6\xa7\x79\xe0\x1a",
.rlen = 48 + 32,
}, { /* RFC 3602 Case 4 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x56\xe4\x7a\x38\xc5\x59\x89\x74"
"\xbc\x46\x90\x3d\xba\x29\x03\x49",
.klen = 8 + 32 + 16,
.iv = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.assoc = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.alen = 16,
.input = "\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf",
.ilen = 64,
.result = "\xc3\x0e\x32\xff\xed\xc0\x77\x4e"
"\x6a\xff\x6a\xf0\x86\x9f\x71\xaa"
"\x0f\x3a\xf0\x7a\x9a\x31\xa9\xc6"
"\x84\xdb\x20\x7e\xb0\xef\x8e\x4e"
"\x35\x90\x7a\xa6\x32\xc3\xff\xdf"
"\x86\x8b\xb7\xb2\x9d\x3d\x46\xad"
"\x83\xce\x9f\x9a\x10\x2e\xe9\x9d"
"\x49\xa5\x3e\x87\xf4\xc3\xda\x55"
"\x7a\x1b\xd4\x3c\xdb\x17\x95\xe2"
"\xe0\x93\xec\xc9\x9f\xf7\xce\xd8"
"\x3f\x54\xe2\x49\x39\xe3\x71\x25"
"\x2b\x6c\xe9\x5d\xec\xec\x2b\x64",
.rlen = 64 + 32,
}, { /* RFC 3602 Case 5 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x90\xd3\x82\xb4\x10\xee\xba\x7a"
"\xd9\x38\xc4\x6c\xec\x1a\x82\xbf",
.klen = 8 + 32 + 16,
.iv = "\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.alen = 24,
.input = "\x08\x00\x0e\xbd\xa7\x0a\x00\x00"
"\x8e\x9c\x08\x3d\xb9\x5b\x07\x00"
crypto: speck - add test vectors for Speck128-XTS Add test vectors for Speck128-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors. Both xts(speck128-generic) and xts-speck128-neon pass these tests. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:22 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0e\x01",
.ilen = 80,
.result = "\xf6\x63\xc2\x5d\x32\x5c\x18\xc6"
"\xa9\x45\x3e\x19\x4e\x12\x08\x49"
"\xa4\x87\x0b\x66\xcc\x6b\x99\x65"
"\x33\x00\x13\xb4\x89\x8d\xc8\x56"
"\xa4\x69\x9e\x52\x3a\x55\xdb\x08"
"\x0b\x59\xec\x3a\x8e\x4b\x7e\x52"
"\x77\x5b\x07\xd1\xdb\x34\xed\x9c"
"\x53\x8a\xb5\x0c\x55\x1b\x87\x4a"
"\xa2\x69\xad\xd0\x47\xad\x2d\x59"
"\x13\xac\x19\xb7\xcf\xba\xd4\xa6"
"\xbb\xd4\x0f\xbe\xa3\x3b\x4c\xb8"
"\x3a\xd2\xe1\x03\x86\xa5\x59\xb7"
"\x73\xc3\x46\x20\x2c\xb1\xef\x68"
"\xbb\x8a\x32\x7e\x12\x8c\x69\xcf",
.rlen = 80 + 32,
}, { /* NIST SP800-38A F.2.3 CBC-AES192.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x8e\x73\xb0\xf7\xda\x0e\x64\x52"
"\xc8\x10\xf3\x2b\x80\x90\x79\xe5"
"\x62\xf8\xea\xd2\x52\x2c\x6b\x7b",
.klen = 8 + 32 + 24,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\x4f\x02\x1d\xb2\x43\xbc\x63\x3d"
"\x71\x78\x18\x3a\x9f\xa0\x71\xe8"
"\xb4\xd9\xad\xa9\xad\x7d\xed\xf4"
"\xe5\xe7\x38\x76\x3f\x69\x14\x5a"
"\x57\x1b\x24\x20\x12\xfb\x7a\xe0"
"\x7f\xa9\xba\xac\x3d\xf1\x02\xe0"
"\x08\xb0\xe2\x79\x88\x59\x88\x81"
"\xd9\x20\xa9\xe6\x4f\x56\x15\xcd"
"\x2f\xee\x5f\xdb\x66\xfe\x79\x09"
"\x61\x81\x31\xea\x5b\x3d\x8e\xfb"
"\xca\x71\x85\x93\xf7\x85\x55\x8b"
"\x7a\xe4\x94\xca\x8b\xba\x19\x33",
.rlen = 64 + 32,
}, { /* NIST SP800-38A F.2.5 CBC-AES256.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x20" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.klen = 8 + 32 + 32,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\xf5\x8c\x4c\x04\xd6\xe5\xf1\xba"
"\x77\x9e\xab\xfb\x5f\x7b\xfb\xd6"
"\x9c\xfc\x4e\x96\x7e\xdb\x80\x8d"
"\x67\x9f\x77\x7b\xc6\x70\x2c\x7d"
"\x39\xf2\x33\x69\xa9\xd9\xba\xcf"
"\xa5\x30\xe2\x63\x04\x23\x14\x61"
"\xb2\xeb\x05\xe2\xc3\x9b\xe9\xfc"
"\xda\x6c\x19\x07\x8c\x6a\x9d\x1b"
"\x24\x29\xed\xc2\x31\x49\xdb\xb1"
"\x8f\x74\xbd\x17\x92\x03\xbe\x8f"
"\xf3\x61\xde\x1c\xe9\xdb\xcd\xd0"
"\xcc\xce\xe9\x85\x57\xcf\x6f\x5f",
.rlen = 64 + 32,
crypto: speck - add support for the Speck block cipher Add a generic implementation of Speck, including the Speck128 and Speck64 variants. Speck is a lightweight block cipher that can be much faster than AES on processors that don't have AES instructions. We are planning to offer Speck-XTS (probably Speck128/256-XTS) as an option for dm-crypt and fscrypt on Android, for low-end mobile devices with older CPUs such as ARMv7 which don't have the Cryptography Extensions. Currently, such devices are unencrypted because AES is not fast enough, even when the NEON bit-sliced implementation of AES is used. Other AES alternatives such as Twofish, Threefish, Camellia, CAST6, and Serpent aren't fast enough either; it seems that only a modern ARX cipher can provide sufficient performance on these devices. This is a replacement for our original proposal (https://patchwork.kernel.org/patch/10101451/) which was to offer ChaCha20 for these devices. However, the use of a stream cipher for disk/file encryption with no space to store nonces would have been much more insecure than we thought initially, given that it would be used on top of flash storage as well as potentially on top of F2FS, neither of which is guaranteed to overwrite data in-place. Speck has been somewhat controversial due to its origin. Nevertheless, it has a straightforward design (it's an ARX cipher), and it appears to be the leading software-optimized lightweight block cipher currently, with the most cryptanalysis. It's also easy to implement without side channels, unlike AES. Moreover, we only intend Speck to be used when the status quo is no encryption, due to AES not being fast enough. We've also considered a novel length-preserving encryption mode based on ChaCha20 and Poly1305. While theoretically attractive, such a mode would be a brand new crypto construction and would be more complicated and difficult to implement efficiently in comparison to Speck-XTS. There is confusion about the byte and word orders of Speck, since the original paper doesn't specify them. But we have implemented it using the orders the authors recommended in a correspondence with them. The test vectors are taken from the original paper but were mapped to byte arrays using the recommended byte and word orders. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:19 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha512_aes_cbc_enc_tv_temp[] = {
{ /* RFC 3602 Case 1 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x06\xa9\x21\x40\x36\xb8\xa1\x5b"
"\x51\x2e\x03\xd5\x34\x12\x00\x06",
.klen = 8 + 64 + 16,
.iv = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.assoc = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.alen = 16,
.input = "Single block msg",
.ilen = 16,
.result = "\xe3\x53\x77\x9c\x10\x79\xae\xb8"
"\x27\x08\x94\x2d\xbe\x77\x18\x1a"
"\x3f\xdc\xad\x90\x03\x63\x5e\x68"
"\xc3\x13\xdd\xa4\x5c\x4d\x54\xa7"
"\x19\x6e\x03\x75\x2b\xa1\x62\xce"
"\xe0\xc6\x96\x75\xb2\x14\xca\x96"
"\xec\xbd\x50\x08\x07\x64\x1a\x49"
"\xe8\x9a\x7c\x06\x3d\xcb\xff\xb2"
"\xfa\x20\x89\xdd\x9c\xac\x9e\x16"
"\x18\x8a\xa0\x6d\x01\x6c\xa3\x3a",
.rlen = 16 + 64,
}, { /* RFC 3602 Case 2 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xc2\x86\x69\x6d\x88\x7c\x9a\xa0"
"\x61\x1b\xbb\x3e\x20\x25\xa4\x5a",
.klen = 8 + 64 + 16,
.iv = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.assoc = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.alen = 16,
.input = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.ilen = 32,
.result = "\xd2\x96\xcd\x94\xc2\xcc\xcf\x8a"
"\x3a\x86\x30\x28\xb5\xe1\xdc\x0a"
"\x75\x86\x60\x2d\x25\x3c\xff\xf9"
"\x1b\x82\x66\xbe\xa6\xd6\x1a\xb1"
"\xda\xb2\x0c\xb2\x26\xc4\xd5\xef"
"\x60\x38\xa4\x5e\x9a\x8c\x1b\x41"
"\x03\x9f\xc4\x64\x7f\x01\x42\x9b"
"\x0e\x1b\xea\xef\xbc\x88\x19\x5e"
"\x31\x7e\xc2\x95\xfc\x09\x32\x0a"
"\x46\x32\x7c\x41\x9c\x59\x3e\xe9"
"\x8f\x9f\xd4\x31\xd6\x22\xbd\xf8"
"\xf7\x0a\x94\xe5\xa9\xc3\xf6\x9d",
.rlen = 32 + 64,
}, { /* RFC 3602 Case 3 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\x6c\x3e\xa0\x47\x76\x30\xce\x21"
"\xa2\xce\x33\x4a\xa7\x46\xc2\xcd",
.klen = 8 + 64 + 16,
.iv = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.assoc = "\xc7\x82\xdc\x4c\x09\x8c\x66\xcb"
"\xd9\xcd\x27\xd8\x25\x68\x2c\x81",
.alen = 16,
.input = "This is a 48-byte message (exactly 3 AES blocks)",
.ilen = 48,
.result = "\xd0\xa0\x2b\x38\x36\x45\x17\x53"
"\xd4\x93\x66\x5d\x33\xf0\xe8\x86"
"\x2d\xea\x54\xcd\xb2\x93\xab\xc7"
"\x50\x69\x39\x27\x67\x72\xf8\xd5"
"\x02\x1c\x19\x21\x6b\xad\x52\x5c"
"\x85\x79\x69\x5d\x83\xba\x26\x84"
"\x64\x19\x17\x5b\x57\xe0\x21\x0f"
"\xca\xdb\xa1\x26\x38\x14\xa2\x69"
"\xdb\x54\x67\x80\xc0\x54\xe0\xfd"
"\x3e\x91\xe7\x91\x7f\x13\x38\x44"
"\xb7\xb1\xd6\xc8\x7d\x48\x8d\x41"
"\x08\xea\x29\x6c\x74\x67\x3f\xb0"
"\xac\x7f\x5c\x1d\xf5\xee\x22\x66"
"\x27\xa6\xb6\x13\xba\xba\xf0\xc2",
.rlen = 48 + 64,
}, { /* RFC 3602 Case 4 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\x56\xe4\x7a\x38\xc5\x59\x89\x74"
"\xbc\x46\x90\x3d\xba\x29\x03\x49",
.klen = 8 + 64 + 16,
.iv = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.assoc = "\x8c\xe8\x2e\xef\xbe\xa0\xda\x3c"
"\x44\x69\x9e\xd7\xdb\x51\xb7\xd9",
.alen = 16,
.input = "\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf",
.ilen = 64,
.result = "\xc3\x0e\x32\xff\xed\xc0\x77\x4e"
"\x6a\xff\x6a\xf0\x86\x9f\x71\xaa"
"\x0f\x3a\xf0\x7a\x9a\x31\xa9\xc6"
"\x84\xdb\x20\x7e\xb0\xef\x8e\x4e"
"\x35\x90\x7a\xa6\x32\xc3\xff\xdf"
"\x86\x8b\xb7\xb2\x9d\x3d\x46\xad"
"\x83\xce\x9f\x9a\x10\x2e\xe9\x9d"
"\x49\xa5\x3e\x87\xf4\xc3\xda\x55"
"\x82\xcd\x42\x28\x21\x20\x15\xcc"
"\xb7\xb2\x48\x40\xc7\x64\x41\x3a"
"\x61\x32\x82\x85\xcf\x27\xed\xb4"
"\xe4\x68\xa2\xf5\x79\x26\x27\xb2"
"\x51\x67\x6a\xc4\xf0\x66\x55\x50"
"\xbc\x6f\xed\xd5\x8d\xde\x23\x7c"
"\x62\x98\x14\xd7\x2f\x37\x8d\xdf"
"\xf4\x33\x80\xeb\x8e\xb4\xa4\xda",
.rlen = 64 + 64,
}, { /* RFC 3602 Case 5 */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x10" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\x90\xd3\x82\xb4\x10\xee\xba\x7a"
"\xd9\x38\xc4\x6c\xec\x1a\x82\xbf",
.klen = 8 + 64 + 16,
.iv = "\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\xe9\x6e\x8c\x08\xab\x46\x57\x63"
"\xfd\x09\x8d\x45\xdd\x3f\xf8\x93",
.alen = 24,
.input = "\x08\x00\x0e\xbd\xa7\x0a\x00\x00"
"\x8e\x9c\x08\x3d\xb9\x5b\x07\x00"
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0e\x01",
.ilen = 80,
.result = "\xf6\x63\xc2\x5d\x32\x5c\x18\xc6"
"\xa9\x45\x3e\x19\x4e\x12\x08\x49"
"\xa4\x87\x0b\x66\xcc\x6b\x99\x65"
"\x33\x00\x13\xb4\x89\x8d\xc8\x56"
"\xa4\x69\x9e\x52\x3a\x55\xdb\x08"
"\x0b\x59\xec\x3a\x8e\x4b\x7e\x52"
"\x77\x5b\x07\xd1\xdb\x34\xed\x9c"
"\x53\x8a\xb5\x0c\x55\x1b\x87\x4a"
"\xa2\x69\xad\xd0\x47\xad\x2d\x59"
"\x13\xac\x19\xb7\xcf\xba\xd4\xa6"
"\x74\x84\x94\xe2\xd7\x7a\xf9\xbf"
"\x00\x8a\xa2\xd5\xb7\xf3\x60\xcf"
"\xa0\x47\xdf\x4e\x09\xf4\xb1\x7f"
"\x14\xd9\x3d\x53\x8e\x12\xb3\x00"
"\x4c\x0a\x4e\x32\x40\x43\x88\xce"
"\x92\x26\xc1\x76\x20\x11\xeb\xba"
"\x62\x4f\x9a\x62\x25\xc3\x75\x80"
"\xb7\x0a\x17\xf5\xd7\x94\xb4\x14",
.rlen = 80 + 64,
}, { /* NIST SP800-38A F.2.3 CBC-AES192.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\x8e\x73\xb0\xf7\xda\x0e\x64\x52"
"\xc8\x10\xf3\x2b\x80\x90\x79\xe5"
"\x62\xf8\xea\xd2\x52\x2c\x6b\x7b",
.klen = 8 + 64 + 24,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\x4f\x02\x1d\xb2\x43\xbc\x63\x3d"
"\x71\x78\x18\x3a\x9f\xa0\x71\xe8"
"\xb4\xd9\xad\xa9\xad\x7d\xed\xf4"
"\xe5\xe7\x38\x76\x3f\x69\x14\x5a"
"\x57\x1b\x24\x20\x12\xfb\x7a\xe0"
"\x7f\xa9\xba\xac\x3d\xf1\x02\xe0"
"\x08\xb0\xe2\x79\x88\x59\x88\x81"
"\xd9\x20\xa9\xe6\x4f\x56\x15\xcd"
"\x77\x4b\x69\x9d\x3a\x0d\xb4\x99"
"\x8f\xc6\x8e\x0e\x72\x58\xe3\x56"
"\xbb\x21\xd2\x7d\x93\x11\x17\x91"
"\xc4\x83\xfd\x0a\xea\x71\xfe\x77"
"\xae\x6f\x0a\xa5\xf0\xcf\xe1\x35"
"\xba\x03\xd5\x32\xfa\x5f\x41\x58"
"\x8d\x43\x98\xa7\x94\x16\x07\x02"
"\x0f\xb6\x81\x50\x28\x95\x2e\x75",
.rlen = 64 + 64,
}, { /* NIST SP800-38A F.2.5 CBC-AES256.Encrypt */
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x20" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
.klen = 8 + 64 + 32,
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.alen = 16,
.input = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ilen = 64,
.result = "\xf5\x8c\x4c\x04\xd6\xe5\xf1\xba"
"\x77\x9e\xab\xfb\x5f\x7b\xfb\xd6"
"\x9c\xfc\x4e\x96\x7e\xdb\x80\x8d"
"\x67\x9f\x77\x7b\xc6\x70\x2c\x7d"
"\x39\xf2\x33\x69\xa9\xd9\xba\xcf"
"\xa5\x30\xe2\x63\x04\x23\x14\x61"
"\xb2\xeb\x05\xe2\xc3\x9b\xe9\xfc"
"\xda\x6c\x19\x07\x8c\x6a\x9d\x1b"
"\xb2\x27\x69\x7f\x45\x64\x79\x2b"
"\xb7\xb8\x4c\xd4\x75\x94\x68\x40"
"\x2a\xea\x91\xc7\x3f\x7c\xed\x7b"
"\x95\x2c\x9b\xa8\xf5\xe5\x52\x8d"
"\x6b\xe1\xae\xf1\x74\xfa\x0d\x0c"
"\xe3\x8d\x64\xc3\x8d\xff\x7c\x8c"
"\xdb\xbf\xa0\xb4\x01\xa2\xa8\xa2"
"\x2c\xb1\x62\x2c\x10\xca\xf1\x21",
.rlen = 64 + 64,
},
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha1_des_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x08" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24",
.klen = 8 + 20 + 8,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x70\xd6\xde\x64\x87\x17\xf1\xe8"
"\x54\x31\x85\x37\xed\x6b\x01\x8d"
"\xe3\xcc\xe0\x1d\x5e\xf3\xfe\xf1"
"\x41\xaa\x33\x91\xa7\x7d\x99\x88"
"\x4d\x85\x6e\x2f\xa3\x69\xf5\x82"
"\x3a\x6f\x25\xcb\x7d\x58\x1f\x9b"
"\xaa\x9c\x11\xd5\x76\x67\xce\xde"
"\x56\xd7\x5a\x80\x69\xea\x3a\x02"
"\xf0\xc7\x7c\xe3\xcb\x40\xe5\x52"
"\xd1\x10\x92\x78\x0b\x8e\x5b\xf1"
"\xe3\x26\x1f\xe1\x15\x41\xc7\xba"
"\x99\xdb\x08\x51\x1c\xd3\x01\xf4"
"\x87\x47\x39\xb8\xd2\xdd\xbd\xfb"
"\x66\x13\xdf\x1c\x01\x44\xf0\x7a"
"\x1a\x6b\x13\xf5\xd5\x0b\xb8\xba"
"\x53\xba\xe1\x76\xe3\x82\x07\x86"
"\x95\x16\x20\x09\xf5\x95\x19\xfd"
"\x3c\xc7\xe0\x42\xc0\x14\x69\xfa"
"\x5c\x44\xa9\x37",
.rlen = 128 + 20,
},
crypto: speck - add test vectors for Speck64-XTS Add test vectors for Speck64-XTS, generated in userspace using C code. The inputs were borrowed from the AES-XTS test vectors, with key lengths adjusted. xts-speck64-neon passes these tests. However, they aren't currently applicable for the generic XTS template, as that only supports a 128-bit block size. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-02-14 11:42:23 -07:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha224_des_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x08" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24",
.klen = 8 + 24 + 8,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x70\xd6\xde\x64\x87\x17\xf1\xe8"
"\x54\x31\x85\x37\xed\x6b\x01\x8d"
"\xe3\xcc\xe0\x1d\x5e\xf3\xfe\xf1"
"\x41\xaa\x33\x91\xa7\x7d\x99\x88"
"\x4d\x85\x6e\x2f\xa3\x69\xf5\x82"
"\x3a\x6f\x25\xcb\x7d\x58\x1f\x9b"
"\xaa\x9c\x11\xd5\x76\x67\xce\xde"
"\x56\xd7\x5a\x80\x69\xea\x3a\x02"
"\xf0\xc7\x7c\xe3\xcb\x40\xe5\x52"
"\xd1\x10\x92\x78\x0b\x8e\x5b\xf1"
"\xe3\x26\x1f\xe1\x15\x41\xc7\xba"
"\x99\xdb\x08\x51\x1c\xd3\x01\xf4"
"\x87\x47\x39\xb8\xd2\xdd\xbd\xfb"
"\x66\x13\xdf\x1c\x01\x44\xf0\x7a"
"\x1a\x6b\x13\xf5\xd5\x0b\xb8\xba"
"\x53\xba\xe1\x76\xe3\x82\x07\x86"
"\x9c\x2d\x7e\xee\x20\x34\x55\x0a"
"\xce\xb5\x4e\x64\x53\xe7\xbf\x91"
"\xab\xd4\xd9\xda\xc9\x12\xae\xf7",
.rlen = 128 + 24,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha256_des_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x08" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24",
.klen = 8 + 32 + 8,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x70\xd6\xde\x64\x87\x17\xf1\xe8"
"\x54\x31\x85\x37\xed\x6b\x01\x8d"
"\xe3\xcc\xe0\x1d\x5e\xf3\xfe\xf1"
"\x41\xaa\x33\x91\xa7\x7d\x99\x88"
"\x4d\x85\x6e\x2f\xa3\x69\xf5\x82"
"\x3a\x6f\x25\xcb\x7d\x58\x1f\x9b"
"\xaa\x9c\x11\xd5\x76\x67\xce\xde"
"\x56\xd7\x5a\x80\x69\xea\x3a\x02"
"\xf0\xc7\x7c\xe3\xcb\x40\xe5\x52"
"\xd1\x10\x92\x78\x0b\x8e\x5b\xf1"
"\xe3\x26\x1f\xe1\x15\x41\xc7\xba"
"\x99\xdb\x08\x51\x1c\xd3\x01\xf4"
"\x87\x47\x39\xb8\xd2\xdd\xbd\xfb"
"\x66\x13\xdf\x1c\x01\x44\xf0\x7a"
"\x1a\x6b\x13\xf5\xd5\x0b\xb8\xba"
"\x53\xba\xe1\x76\xe3\x82\x07\x86"
"\xc6\x58\xa1\x60\x70\x91\x39\x36"
"\x50\xf6\x5d\xab\x4b\x51\x4e\x5e"
"\xde\x63\xde\x76\x52\xde\x9f\xba"
"\x90\xcf\x15\xf2\xbb\x6e\x84\x00",
.rlen = 128 + 32,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha384_des_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x08" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24",
.klen = 8 + 48 + 8,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x70\xd6\xde\x64\x87\x17\xf1\xe8"
"\x54\x31\x85\x37\xed\x6b\x01\x8d"
"\xe3\xcc\xe0\x1d\x5e\xf3\xfe\xf1"
"\x41\xaa\x33\x91\xa7\x7d\x99\x88"
"\x4d\x85\x6e\x2f\xa3\x69\xf5\x82"
"\x3a\x6f\x25\xcb\x7d\x58\x1f\x9b"
"\xaa\x9c\x11\xd5\x76\x67\xce\xde"
"\x56\xd7\x5a\x80\x69\xea\x3a\x02"
"\xf0\xc7\x7c\xe3\xcb\x40\xe5\x52"
"\xd1\x10\x92\x78\x0b\x8e\x5b\xf1"
"\xe3\x26\x1f\xe1\x15\x41\xc7\xba"
"\x99\xdb\x08\x51\x1c\xd3\x01\xf4"
"\x87\x47\x39\xb8\xd2\xdd\xbd\xfb"
"\x66\x13\xdf\x1c\x01\x44\xf0\x7a"
"\x1a\x6b\x13\xf5\xd5\x0b\xb8\xba"
"\x53\xba\xe1\x76\xe3\x82\x07\x86"
"\xa8\x8e\x9c\x74\x8c\x2b\x99\xa0"
"\xc8\x8c\xef\x25\x07\x83\x11\x3a"
"\x31\x8d\xbe\x3b\x6a\xd7\x96\xfe"
"\x5e\x67\xb5\x74\xe7\xe7\x85\x61"
"\x6a\x95\x26\x75\xcc\x53\x89\xf3"
"\x74\xc9\x2a\x76\x20\xa2\x64\x62",
.rlen = 128 + 48,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha512_des_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x08" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24",
.klen = 8 + 64 + 8,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x70\xd6\xde\x64\x87\x17\xf1\xe8"
"\x54\x31\x85\x37\xed\x6b\x01\x8d"
"\xe3\xcc\xe0\x1d\x5e\xf3\xfe\xf1"
"\x41\xaa\x33\x91\xa7\x7d\x99\x88"
"\x4d\x85\x6e\x2f\xa3\x69\xf5\x82"
"\x3a\x6f\x25\xcb\x7d\x58\x1f\x9b"
"\xaa\x9c\x11\xd5\x76\x67\xce\xde"
"\x56\xd7\x5a\x80\x69\xea\x3a\x02"
"\xf0\xc7\x7c\xe3\xcb\x40\xe5\x52"
"\xd1\x10\x92\x78\x0b\x8e\x5b\xf1"
"\xe3\x26\x1f\xe1\x15\x41\xc7\xba"
"\x99\xdb\x08\x51\x1c\xd3\x01\xf4"
"\x87\x47\x39\xb8\xd2\xdd\xbd\xfb"
"\x66\x13\xdf\x1c\x01\x44\xf0\x7a"
"\x1a\x6b\x13\xf5\xd5\x0b\xb8\xba"
"\x53\xba\xe1\x76\xe3\x82\x07\x86"
"\xc6\x2c\x73\x88\xb0\x9d\x5f\x3e"
"\x5b\x78\xca\x0e\xab\x8a\xa3\xbb"
"\xd9\x1d\xc3\xe3\x05\xac\x76\xfb"
"\x58\x83\xda\x67\xfb\x21\x24\xa2"
"\xb1\xa7\xd7\x66\xa6\x8d\xa6\x93"
"\x97\xe2\xe3\xb8\xaa\x48\x85\xee"
"\x8c\xf6\x07\x95\x1f\xa6\x6c\x96"
"\x99\xc7\x5c\x8d\xd8\xb5\x68\x7b",
.rlen = 128 + 64,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha1_des3_ede_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 8 + 20 + 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19"
"\x67\x6d\xb1\xf5\xb8\x10\xdc\xc6"
"\x75\x86\x96\x6b\xb1\xc5\xe4\xcf"
"\xd1\x60\x91\xb3",
.rlen = 128 + 20,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha224_des3_ede_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 8 + 24 + 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19"
"\x15\x24\x7f\x5a\x45\x4a\x66\xce"
"\x2b\x0b\x93\x99\x2f\x9d\x0c\x6c"
"\x56\x1f\xe1\xa6\x41\xb2\x4c\xd0",
.rlen = 128 + 24,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha256_des3_ede_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 8 + 32 + 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19"
"\x73\xb0\xea\x9f\xe8\x18\x80\xd6"
"\x56\x38\x44\xc0\xdb\xe3\x4f\x71"
"\xf7\xce\xd1\xd3\xf8\xbd\x3e\x4f"
"\xca\x43\x95\xdf\x80\x61\x81\xa9",
.rlen = 128 + 32,
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec hmac_sha384_des3_ede_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 8 + 48 + 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19"
"\x6d\x77\xfc\x80\x9d\x8a\x9c\xb7"
"\x70\xe7\x93\xbf\x73\xe6\x9f\x83"
"\x99\x62\x23\xe6\x5b\xd0\xda\x18"
"\xa4\x32\x8a\x0b\x46\xd7\xf0\x39"
"\x36\x5d\x13\x2f\x86\x10\x78\xd6"
"\xd6\xbe\x5c\xb9\x15\x89\xf9\x1b",
.rlen = 128 + 48,
},
};
static const struct aead_testvec hmac_sha512_des3_ede_cbc_enc_tv_temp[] = {
{ /*Generated with cryptopp*/
#ifdef __LITTLE_ENDIAN
.key = "\x08\x00" /* rta length */
"\x01\x00" /* rta type */
#else
.key = "\x00\x08" /* rta length */
"\x00\x01" /* rta type */
#endif
"\x00\x00\x00\x18" /* enc key length */
"\x11\x22\x33\x44\x55\x66\x77\x88"
"\x99\xaa\xbb\xcc\xdd\xee\xff\x11"
"\x22\x33\x44\x55\x66\x77\x88\x99"
"\xaa\xbb\xcc\xdd\xee\xff\x11\x22"
"\x33\x44\x55\x66\x77\x88\x99\xaa"
"\xbb\xcc\xdd\xee\xff\x11\x22\x33"
"\x44\x55\x66\x77\x88\x99\xaa\xbb"
"\xcc\xdd\xee\xff\x11\x22\x33\x44"
"\xE9\xC0\xFF\x2E\x76\x0B\x64\x24"
"\x44\x4D\x99\x5A\x12\xD6\x40\xC0"
"\xEA\xC2\x84\xE8\x14\x95\xDB\xE8",
.klen = 8 + 64 + 24,
.iv = "\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x01"
"\x7D\x33\x88\x93\x0F\x93\xB2\x42",
.alen = 16,
.input = "\x6f\x54\x20\x6f\x61\x4d\x79\x6e"
"\x53\x20\x63\x65\x65\x72\x73\x74"
"\x54\x20\x6f\x6f\x4d\x20\x6e\x61"
"\x20\x79\x65\x53\x72\x63\x74\x65"
"\x20\x73\x6f\x54\x20\x6f\x61\x4d"
"\x79\x6e\x53\x20\x63\x65\x65\x72"
"\x73\x74\x54\x20\x6f\x6f\x4d\x20"
"\x6e\x61\x20\x79\x65\x53\x72\x63"
"\x74\x65\x20\x73\x6f\x54\x20\x6f"
"\x61\x4d\x79\x6e\x53\x20\x63\x65"
"\x65\x72\x73\x74\x54\x20\x6f\x6f"
"\x4d\x20\x6e\x61\x20\x79\x65\x53"
"\x72\x63\x74\x65\x20\x73\x6f\x54"
"\x20\x6f\x61\x4d\x79\x6e\x53\x20"
"\x63\x65\x65\x72\x73\x74\x54\x20"
"\x6f\x6f\x4d\x20\x6e\x61\x0a\x79",
.ilen = 128,
.result = "\x0e\x2d\xb6\x97\x3c\x56\x33\xf4"
"\x67\x17\x21\xc7\x6e\x8a\xd5\x49"
"\x74\xb3\x49\x05\xc5\x1c\xd0\xed"
"\x12\x56\x5c\x53\x96\xb6\x00\x7d"
"\x90\x48\xfc\xf5\x8d\x29\x39\xcc"
"\x8a\xd5\x35\x18\x36\x23\x4e\xd7"
"\x76\xd1\xda\x0c\x94\x67\xbb\x04"
"\x8b\xf2\x03\x6c\xa8\xcf\xb6\xea"
"\x22\x64\x47\xaa\x8f\x75\x13\xbf"
"\x9f\xc2\xc3\xf0\xc9\x56\xc5\x7a"
"\x71\x63\x2e\x89\x7b\x1e\x12\xca"
"\xe2\x5f\xaf\xd8\xa4\xf8\xc9\x7a"
"\xd6\xf9\x21\x31\x62\x44\x45\xa6"
"\xd6\xbc\x5a\xd3\x2d\x54\x43\xcc"
"\x9d\xde\xa5\x70\xe9\x42\x45\x8a"
"\x6b\xfa\xb1\x91\x13\xb0\xd9\x19"
"\x41\xb5\x1f\xbb\xbd\x4e\xb8\x32"
"\x22\x86\x4e\x57\x1b\x2a\xd8\x6e"
"\xa9\xfb\xc8\xf3\xbf\x2d\xae\x2b"
"\x3b\xbc\x41\xe8\x38\xbb\xf1\x60"
"\x4c\x68\xa9\x4e\x8c\x73\xa7\xc0"
"\x2a\x74\xd4\x65\x12\xcb\x55\xf2"
"\xd5\x02\x6d\xe6\xaf\xc9\x2f\xf2"
"\x57\xaa\x85\xf7\xf3\x6a\xcb\xdb",
.rlen = 128 + 64,
},
};
static const struct cipher_testvec aes_lrw_tv_template[] = {
/* from http://grouper.ieee.org/groups/1619/email/pdf00017.pdf */
{ /* LRW-32-AES 1 */
.key = "\x45\x62\xac\x25\xf8\x28\x17\x6d"
"\x4c\x26\x84\x14\xb5\x68\x01\x85"
"\x25\x8e\x2a\x05\xe7\x3e\x9d\x03"
"\xee\x5a\x83\x0c\xcc\x09\x4c\x87",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xf1\xb2\x73\xcd\x65\xa3\xdf\x5f"
"\xe9\x5d\x48\x92\x54\x63\x4e\xb8",
.len = 16,
}, { /* LRW-32-AES 2 */
.key = "\x59\x70\x47\x14\xf5\x57\x47\x8c"
"\xd7\x79\xe8\x0f\x54\x88\x79\x44"
"\x0d\x48\xf0\xb7\xb1\x5a\x53\xea"
"\x1c\xaa\x6b\x29\xc2\xca\xfb\xaf",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x00\xc8\x2b\xae\x95\xbb\xcd\xe5"
"\x27\x4f\x07\x69\xb2\x60\xe1\x36",
.len = 16,
}, { /* LRW-32-AES 3 */
.key = "\xd8\x2a\x91\x34\xb2\x6a\x56\x50"
"\x30\xfe\x69\xe2\x37\x7f\x98\x47"
"\xcd\xf9\x0b\x16\x0c\x64\x8f\xb6"
"\xb0\x0d\x0d\x1b\xae\x85\x87\x1f",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x76\x32\x21\x83\xed\x8f\xf1\x82"
"\xf9\x59\x62\x03\x69\x0e\x5e\x01",
.len = 16,
}, { /* LRW-32-AES 4 */
.key = "\x0f\x6a\xef\xf8\xd3\xd2\xbb\x15"
"\x25\x83\xf7\x3c\x1f\x01\x28\x74"
"\xca\xc6\xbc\x35\x4d\x4a\x65\x54"
"\x90\xae\x61\xcf\x7b\xae\xbd\xcc"
"\xad\xe4\x94\xc5\x4a\x29\xae\x70",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x9c\x0f\x15\x2f\x55\xa2\xd8\xf0"
"\xd6\x7b\x8f\x9e\x28\x22\xbc\x41",
.len = 16,
}, { /* LRW-32-AES 5 */
.key = "\x8a\xd4\xee\x10\x2f\xbd\x81\xff"
"\xf8\x86\xce\xac\x93\xc5\xad\xc6"
"\xa0\x19\x07\xc0\x9d\xf7\xbb\xdd"
"\x52\x13\xb2\xb7\xf0\xff\x11\xd8"
"\xd6\x08\xd0\xcd\x2e\xb1\x17\x6f",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xd4\x27\x6a\x7f\x14\x91\x3d\x65"
"\xc8\x60\x48\x02\x87\xe3\x34\x06",
.len = 16,
}, { /* LRW-32-AES 6 */
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\xbd\x06\xb8\xe1\xdb\x98\x89\x9e"
"\xc4\x98\xe4\x91\xcf\x1c\x70\x2b",
.len = 16,
}, { /* LRW-32-AES 7 */
.key = "\xfb\x76\x15\xb2\x3d\x80\x89\x1d"
"\xd4\x70\x98\x0b\xc7\x95\x84\xc8"
"\xb2\xfb\x64\xce\x60\x97\x87\x8d"
"\x17\xfc\xe4\x5a\x49\xe8\x30\xb7"
"\x6e\x78\x17\xe7\x2d\x5e\x12\xd4"
"\x60\x64\x04\x7a\xf1\x2f\x9e\x0c",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x5b\x90\x8e\xc1\xab\xdd\x67\x5f"
"\x3d\x69\x8a\x95\x53\xc8\x9c\xe5",
.len = 16,
}, {
/* http://www.mail-archive.com/stds-p1619@listserv.ieee.org/msg00173.html */
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x05\x11\xb7\x18\xab\xc6\x2d\xac"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x70\x5d\xf6\x22\x94\xcd\xe5\x6c"
"\x17\x6b\xf6\x1c\xf0\xf3\x6e\xf8"
"\x50\x38\x1f\x71\x49\xb6\x57\xd6"
"\x8f\xcb\x8d\x6b\xe3\xa6\x29\x90"
"\xfe\x2a\x62\x82\xae\x6d\x8b\xf6"
"\xad\x1e\x9e\x20\x5f\x38\xbe\x04"
"\xda\x10\x8e\xed\xa2\xa4\x87\xab"
"\xda\x6b\xb4\x0c\x75\xba\xd3\x7c"
"\xc9\xac\x42\x31\x95\x7c\xc9\x04"
"\xeb\xd5\x6e\x32\x69\x8a\xdb\xa6"
"\x15\xd7\x3f\x4f\x2f\x66\x69\x03"
"\x9c\x1f\x54\x0f\xde\x1f\xf3\x65"
"\x4c\x96\x12\xed\x7c\x92\x03\x01"
"\x6f\xbc\x35\x93\xac\xf1\x27\xf1"
"\xb4\x96\x82\x5a\x5f\xb0\xa0\x50"
"\x89\xa4\x8e\x66\x44\x85\xcc\xfd"
"\x33\x14\x70\xe3\x96\xb2\xc3\xd3"
"\xbb\x54\x5a\x1a\xf9\x74\xa2\xc5"
"\x2d\x64\x75\xdd\xb4\x54\xe6\x74"
"\x8c\xd3\x9d\x9e\x86\xab\x51\x53"
"\xb7\x93\x3e\x6f\xd0\x4e\x2c\x40"
"\xf6\xa8\x2e\x3e\x9d\xf4\x66\xa5"
"\x76\x12\x73\x44\x1a\x56\xd7\x72"
"\x88\xcd\x21\x8c\x4c\x0f\xfe\xda"
"\x95\xe0\x3a\xa6\xa5\x84\x46\xcd"
"\xd5\x3e\x9d\x3a\xe2\x67\xe6\x60"
"\x1a\xe2\x70\x85\x58\xc2\x1b\x09"
"\xe1\xd7\x2c\xca\xad\xa8\x8f\xf9"
"\xac\xb3\x0e\xdb\xca\x2e\xe2\xb8"
"\x51\x71\xd9\x3c\x6c\xf1\x56\xf8"
"\xea\x9c\xf1\xfb\x0c\xe6\xb7\x10"
"\x1c\xf8\xa9\x7c\xe8\x53\x35\xc1"
"\x90\x3e\x76\x4a\x74\xa4\x21\x2c"
"\xf6\x2c\x4e\x0f\x94\x3a\x88\x2e"
"\x41\x09\x6a\x33\x7d\xf6\xdd\x3f"
"\x8d\x23\x31\x74\x84\xeb\x88\x6e"
"\xcc\xb9\xbc\x22\x83\x19\x07\x22"
"\xa5\x2d\xdf\xa5\xf3\x80\x85\x78"
"\x84\x39\x6a\x6d\x6a\x99\x4f\xa5"
"\x15\xfe\x46\xb0\xe4\x6c\xa5\x41"
"\x3c\xce\x8f\x42\x60\x71\xa7\x75"
"\x08\x40\x65\x8a\x82\xbf\xf5\x43"
"\x71\x96\xa9\x4d\x44\x8a\x20\xbe"
"\xfa\x4d\xbb\xc0\x7d\x31\x96\x65"
"\xe7\x75\xe5\x3e\xfd\x92\x3b\xc9"
"\x55\xbb\x16\x7e\xf7\xc2\x8c\xa4"
"\x40\x1d\xe5\xef\x0e\xdf\xe4\x9a"
"\x62\x73\x65\xfd\x46\x63\x25\x3d"
"\x2b\xaf\xe5\x64\xfe\xa5\x5c\xcf"
"\x24\xf3\xb4\xac\x64\xba\xdf\x4b"
"\xc6\x96\x7d\x81\x2d\x8d\x97\xf7"
"\xc5\x68\x77\x84\x32\x2b\xcc\x85"
"\x74\x96\xf0\x12\x77\x61\xb9\xeb"
"\x71\xaa\x82\xcb\x1c\xdb\x89\xc8"
"\xc6\xb5\xe3\x5c\x7d\x39\x07\x24"
"\xda\x39\x87\x45\xc0\x2b\xbb\x01"
"\xac\xbc\x2a\x5c\x7f\xfc\xe8\xce"
"\x6d\x9c\x6f\xed\xd3\xc1\xa1\xd6"
"\xc5\x55\xa9\x66\x2f\xe1\xc8\x32"
"\xa6\x5d\xa4\x3a\x98\x73\xe8\x45"
"\xa4\xc7\xa8\xb4\xf6\x13\x03\xf6"
"\xe9\x2e\xc4\x29\x0f\x84\xdb\xc4"
"\x21\xc4\xc2\x75\x67\x89\x37\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x1a\x1d\xa9\x30\xad\xf9\x2f\x9b"
"\xb6\x1d\xae\xef\xf0\x2f\xf8\x5a"
"\x39\x3c\xbf\x2a\xb2\x45\xb2\x23"
"\x1b\x63\x3c\xcf\xaa\xbe\xcf\x4e"
"\xfa\xe8\x29\xc2\x20\x68\x2b\x3c"
"\x2e\x8b\xf7\x6e\x25\xbd\xe3\x3d"
"\x66\x27\xd6\xaf\xd6\x64\x3e\xe3"
"\xe8\x58\x46\x97\x39\x51\x07\xde"
"\xcb\x37\xbc\xa9\xc0\x5f\x75\xc3"
"\x0e\x84\x23\x1d\x16\xd4\x1c\x59"
"\x9c\x1a\x02\x55\xab\x3a\x97\x1d"
"\xdf\xdd\xc7\x06\x51\xd7\x70\xae"
"\x23\xc6\x8c\xf5\x1e\xa0\xe5\x82"
"\xb8\xb2\xbf\x04\xa0\x32\x8e\x68"
"\xeb\xaf\x6e\x2d\x94\x22\x2f\xce"
"\x4c\xb5\x59\xe2\xa2\x2f\xa0\x98"
"\x1a\x97\xc6\xd4\xb5\x00\x59\xf2"
"\x84\x14\x72\xb1\x9a\x6e\xa3\x7f"
"\xea\x20\xe7\xcb\x65\x77\x3a\xdf"
"\xc8\x97\x67\x15\xc2\x2a\x27\xcc"
"\x18\x55\xa1\x24\x0b\x24\x24\xaf"
"\x5b\xec\x68\xb8\xc8\xf5\xba\x63"
"\xff\xed\x89\xce\xd5\x3d\x88\xf3"
"\x25\xef\x05\x7c\x3a\xef\xeb\xd8"
"\x7a\x32\x0d\xd1\x1e\x58\x59\x99"
"\x90\x25\xb5\x26\xb0\xe3\x2b\x6c"
"\x4c\xa9\x8b\x84\x4f\x5e\x01\x50"
"\x41\x30\x58\xc5\x62\x74\x52\x1d"
"\x45\x24\x6a\x42\x64\x4f\x97\x1c"
"\xa8\x66\xb5\x6d\x79\xd4\x0d\x48"
"\xc5\x5f\xf3\x90\x32\xdd\xdd\xe1"
"\xe4\xa9\x9f\xfc\xc3\x52\x5a\x46"
"\xe4\x81\x84\x95\x36\x59\x7a\x6b"
"\xaa\xb3\x60\xad\xce\x9f\x9f\x28"
"\xe0\x01\x75\x22\xc4\x4e\xa9\x62"
"\x5c\x62\x0d\x00\xcb\x13\xe8\x43"
"\x72\xd4\x2d\x53\x46\xb5\xd1\x16"
"\x22\x18\xdf\x34\x33\xf5\xd6\x1c"
"\xb8\x79\x78\x97\x94\xff\x72\x13"
"\x4c\x27\xfc\xcb\xbf\x01\x53\xa6"
"\xb4\x50\x6e\xde\xdf\xb5\x43\xa4"
"\x59\xdf\x52\xf9\x7c\xe0\x11\x6f"
"\x2d\x14\x8e\x24\x61\x2c\xe1\x17"
"\xcc\xce\x51\x0c\x19\x8a\x82\x30"
"\x94\xd5\x3d\x6a\x53\x06\x5e\xbd"
"\xb7\xeb\xfa\xfd\x27\x51\xde\x85"
"\x1e\x86\x53\x11\x53\x94\x00\xee"
"\x2b\x8c\x08\x2a\xbf\xdd\xae\x11"
"\xcb\x1e\xa2\x07\x9a\x80\xcf\x62"
"\x9b\x09\xdc\x95\x3c\x96\x8e\xb1"
"\x09\xbd\xe4\xeb\xdb\xca\x70\x7a"
"\x9e\xfa\x31\x18\x45\x3c\x21\x33"
"\xb0\xb3\x2b\xea\xf3\x71\x2d\xe1"
"\x03\xad\x1b\x48\xd4\x67\x27\xf0"
"\x62\xe4\x3d\xfb\x9b\x08\x76\xe7"
"\xdd\x2b\x01\x39\x04\x5a\x58\x7a"
"\xf7\x11\x90\xec\xbd\x51\x5c\x32"
"\x6b\xd7\x35\x39\x02\x6b\xf2\xa6"
"\xd0\x0d\x07\xe1\x06\xc4\x5b\x7d"
"\xe4\x6a\xd7\xee\x15\x1f\x83\xb4"
"\xa3\xa7\x5e\xc3\x90\xb7\xef\xd3"
"\xb7\x4f\xf8\x92\x4c\xb7\x3c\x29"
"\xcd\x7e\x2b\x5d\x43\xea\x42\xe7"
"\x74\x3f\x7d\x58\x88\x75\xde\x3e",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec aes_xts_tv_template[] = {
/* http://grouper.ieee.org/groups/1619/email/pdf00086.pdf */
{ /* XTS-AES 1 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.fips_skip = 1,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x91\x7c\xf6\x9e\xbd\x68\xb2\xec"
"\x9b\x9f\xe9\xa3\xea\xdd\xa6\x92"
"\xcd\x43\xd2\xf5\x95\x98\xed\x85"
"\x8c\x02\xc2\x65\x2f\xbf\x92\x2e",
.len = 32,
}, { /* XTS-AES 2 */
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\xc4\x54\x18\x5e\x6a\x16\x93\x6e"
"\x39\x33\x40\x38\xac\xef\x83\x8b"
"\xfb\x18\x6f\xff\x74\x80\xad\xc4"
"\x28\x93\x82\xec\xd6\xd3\x94\xf0",
.len = 32,
}, { /* XTS-AES 3 */
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
.ctext = "\xaf\x85\x33\x6b\x59\x7a\xfc\x1a"
"\x90\x0b\x2e\xb2\x1e\xc9\x49\xd2"
"\x92\xdf\x4c\x04\x7e\x0b\x21\x53"
"\x21\x86\xa5\x97\x1a\x22\x7a\x89",
.len = 32,
}, { /* XTS-AES 4 */
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x23\x84\x62\x64\x33\x83\x27\x95",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x27\xa7\x47\x9b\xef\xa1\xd4\x76"
"\x48\x9f\x30\x8c\xd4\xcf\xa6\xe2"
"\xa9\x6e\x4b\xbe\x32\x08\xff\x25"
"\x28\x7d\xd3\x81\x96\x16\xe8\x9c"
"\xc7\x8c\xf7\xf5\xe5\x43\x44\x5f"
"\x83\x33\xd8\xfa\x7f\x56\x00\x00"
"\x05\x27\x9f\xa5\xd8\xb5\xe4\xad"
"\x40\xe7\x36\xdd\xb4\xd3\x54\x12"
"\x32\x80\x63\xfd\x2a\xab\x53\xe5"
"\xea\x1e\x0a\x9f\x33\x25\x00\xa5"
"\xdf\x94\x87\xd0\x7a\x5c\x92\xcc"
"\x51\x2c\x88\x66\xc7\xe8\x60\xce"
"\x93\xfd\xf1\x66\xa2\x49\x12\xb4"
"\x22\x97\x61\x46\xae\x20\xce\x84"
"\x6b\xb7\xdc\x9b\xa9\x4a\x76\x7a"
"\xae\xf2\x0c\x0d\x61\xad\x02\x65"
"\x5e\xa9\x2d\xc4\xc4\xe4\x1a\x89"
"\x52\xc6\x51\xd3\x31\x74\xbe\x51"
"\xa1\x0c\x42\x11\x10\xe6\xd8\x15"
"\x88\xed\xe8\x21\x03\xa2\x52\xd8"
"\xa7\x50\xe8\x76\x8d\xef\xff\xed"
"\x91\x22\x81\x0a\xae\xb9\x9f\x91"
"\x72\xaf\x82\xb6\x04\xdc\x4b\x8e"
"\x51\xbc\xb0\x82\x35\xa6\xf4\x34"
"\x13\x32\xe4\xca\x60\x48\x2a\x4b"
"\xa1\xa0\x3b\x3e\x65\x00\x8f\xc5"
"\xda\x76\xb7\x0b\xf1\x69\x0d\xb4"
"\xea\xe2\x9c\x5f\x1b\xad\xd0\x3c"
"\x5c\xcf\x2a\x55\xd7\x05\xdd\xcd"
"\x86\xd4\x49\x51\x1c\xeb\x7e\xc3"
"\x0b\xf1\x2b\x1f\xa3\x5b\x91\x3f"
"\x9f\x74\x7a\x8a\xfd\x1b\x13\x0e"
"\x94\xbf\xf9\x4e\xff\xd0\x1a\x91"
"\x73\x5c\xa1\x72\x6a\xcd\x0b\x19"
"\x7c\x4e\x5b\x03\x39\x36\x97\xe1"
"\x26\x82\x6f\xb6\xbb\xde\x8e\xcc"
"\x1e\x08\x29\x85\x16\xe2\xc9\xed"
"\x03\xff\x3c\x1b\x78\x60\xf6\xde"
"\x76\xd4\xce\xcd\x94\xc8\x11\x98"
"\x55\xef\x52\x97\xca\x67\xe9\xf3"
"\xe7\xff\x72\xb1\xe9\x97\x85\xca"
"\x0a\x7e\x77\x20\xc5\xb3\x6d\xc6"
"\xd7\x2c\xac\x95\x74\xc8\xcb\xbc"
"\x2f\x80\x1e\x23\xe5\x6f\xd3\x44"
"\xb0\x7f\x22\x15\x4b\xeb\xa0\xf0"
"\x8c\xe8\x89\x1e\x64\x3e\xd9\x95"
"\xc9\x4d\x9a\x69\xc9\xf1\xb5\xf4"
"\x99\x02\x7a\x78\x57\x2a\xee\xbd"
"\x74\xd2\x0c\xc3\x98\x81\xc2\x13"
"\xee\x77\x0b\x10\x10\xe4\xbe\xa7"
"\x18\x84\x69\x77\xae\x11\x9f\x7a"
"\x02\x3a\xb5\x8c\xca\x0a\xd7\x52"
"\xaf\xe6\x56\xbb\x3c\x17\x25\x6a"
"\x9f\x6e\x9b\xf1\x9f\xdd\x5a\x38"
"\xfc\x82\xbb\xe8\x72\xc5\x53\x9e"
"\xdb\x60\x9e\xf4\xf7\x9c\x20\x3e"
"\xbb\x14\x0f\x2e\x58\x3c\xb2\xad"
"\x15\xb4\xaa\x5b\x65\x50\x16\xa8"
"\x44\x92\x77\xdb\xd4\x77\xef\x2c"
"\x8d\x6c\x01\x7d\xb7\x38\xb1\x8d"
"\xeb\x4a\x42\x7d\x19\x23\xce\x3f"
"\xf2\x62\x73\x57\x79\xa4\x18\xf2"
"\x0a\x28\x2d\xf9\x20\x14\x7b\xea"
"\xbe\x42\x1e\xe5\x31\x9d\x05\x68",
.len = 512,
}, { /* XTS-AES 10, XTS-AES-256, data unit 512 bytes */
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add larger cast6 testvectors New ECB, CBC, CTR, LRW and XTS testvectors for cast6. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:38:29 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x1c\x3b\x3a\x10\x2f\x77\x03\x86"
"\xe4\x83\x6c\x99\xe3\x70\xcf\x9b"
"\xea\x00\x80\x3f\x5e\x48\x23\x57"
"\xa4\xae\x12\xd4\x14\xa3\xe6\x3b"
"\x5d\x31\xe2\x76\xf8\xfe\x4a\x8d"
"\x66\xb3\x17\xf9\xac\x68\x3f\x44"
"\x68\x0a\x86\xac\x35\xad\xfc\x33"
"\x45\xbe\xfe\xcb\x4b\xb1\x88\xfd"
"\x57\x76\x92\x6c\x49\xa3\x09\x5e"
"\xb1\x08\xfd\x10\x98\xba\xec\x70"
"\xaa\xa6\x69\x99\xa7\x2a\x82\xf2"
"\x7d\x84\x8b\x21\xd4\xa7\x41\xb0"
"\xc5\xcd\x4d\x5f\xff\x9d\xac\x89"
"\xae\xba\x12\x29\x61\xd0\x3a\x75"
"\x71\x23\xe9\x87\x0f\x8a\xcf\x10"
"\x00\x02\x08\x87\x89\x14\x29\xca"
"\x2a\x3e\x7a\x7d\x7d\xf7\xb1\x03"
"\x55\x16\x5c\x8b\x9a\x6d\x0a\x7d"
"\xe8\xb0\x62\xc4\x50\x0d\xc4\xcd"
"\x12\x0c\x0f\x74\x18\xda\xe3\xd0"
"\xb5\x78\x1c\x34\x80\x3f\xa7\x54"
"\x21\xc7\x90\xdf\xe1\xde\x18\x34"
"\xf2\x80\xd7\x66\x7b\x32\x7f\x6c"
"\x8c\xd7\x55\x7e\x12\xac\x3a\x0f"
"\x93\xec\x05\xc5\x2e\x04\x93\xef"
"\x31\xa1\x2d\x3d\x92\x60\xf7\x9a"
"\x28\x9d\x6a\x37\x9b\xc7\x0c\x50"
"\x84\x14\x73\xd1\xa8\xcc\x81\xec"
"\x58\x3e\x96\x45\xe0\x7b\x8d\x96"
"\x70\x65\x5b\xa5\xbb\xcf\xec\xc6"
"\xdc\x39\x66\x38\x0a\xd8\xfe\xcb"
"\x17\xb6\xba\x02\x46\x9a\x02\x0a"
"\x84\xe1\x8e\x8f\x84\x25\x20\x70"
"\xc1\x3e\x9f\x1f\x28\x9b\xe5\x4f"
"\xbc\x48\x14\x57\x77\x8f\x61\x60"
"\x15\xe1\x32\x7a\x02\xb1\x40\xf1"
"\x50\x5e\xb3\x09\x32\x6d\x68\x37"
"\x8f\x83\x74\x59\x5c\x84\x9d\x84"
"\xf4\xc3\x33\xec\x44\x23\x88\x51"
"\x43\xcb\x47\xbd\x71\xc5\xed\xae"
"\x9b\xe6\x9a\x2f\xfe\xce\xb1\xbe"
"\xc9\xde\x24\x4f\xbe\x15\x99\x2b"
"\x11\xb7\x7c\x04\x0f\x12\xbd\x8f"
"\x6a\x97\x5a\x44\xa0\xf9\x0c\x29"
"\xa9\xab\xc3\xd4\xd8\x93\x92\x72"
"\x84\xc5\x87\x54\xcc\xe2\x94\x52"
"\x9f\x86\x14\xdc\xd2\xab\xa9\x91"
"\x92\x5f\xed\xc4\xae\x74\xff\xac"
"\x6e\x33\x3b\x93\xeb\x4a\xff\x04"
"\x79\xda\x9a\x41\x0e\x44\x50\xe0"
"\xdd\x7a\xe4\xc6\xe2\x91\x09\x00"
"\x57\x5d\xa4\x01\xfc\x07\x05\x9f"
"\x64\x5e\x8b\x7e\x9b\xfd\xef\x33"
"\x94\x30\x54\xff\x84\x01\x14\x93"
"\xc2\x7b\x34\x29\xea\xed\xb4\xed"
"\x53\x76\x44\x1a\x77\xed\x43\x85"
"\x1a\xd7\x7f\x16\xf5\x41\xdf\xd2"
"\x69\xd5\x0d\x6a\x5f\x14\xfb\x0a"
"\xab\x1c\xbb\x4c\x15\x50\xbe\x97"
"\xf7\xab\x40\x66\x19\x3c\x4c\xaa"
"\x77\x3d\xad\x38\x01\x4b\xd2\x09"
"\x2f\xa7\x55\xc8\x24\xbb\x5e\x54"
"\xc4\xf3\x6f\xfd\xa9\xfc\xea\x70"
"\xb9\xc6\xe6\x93\xe1\x48\xc1\x51",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec aes_ctr_tv_template[] = {
{ /* From NIST Special Publication 800-38A, Appendix F.5 */
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\x87\x4d\x61\x91\xb6\x20\xe3\x26"
"\x1b\xef\x68\x64\x99\x0d\xb6\xce"
"\x98\x06\xf6\x6b\x79\x70\xfd\xff"
"\x86\x17\x18\x7b\xb9\xff\xfd\xff"
"\x5a\xe4\xdf\x3e\xdb\xd5\xd3\x5e"
"\x5b\x4f\x09\x02\x0d\xb0\x3e\xab"
"\x1e\x03\x1d\xda\x2f\xbe\x03\xd1"
"\x79\x21\x70\xa0\xf3\x00\x9c\xee",
.len = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8e\x73\xb0\xf7\xda\x0e\x64\x52"
"\xc8\x10\xf3\x2b\x80\x90\x79\xe5"
"\x62\xf8\xea\xd2\x52\x2c\x6b\x7b",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\x1a\xbc\x93\x24\x17\x52\x1c\xa2"
"\x4f\x2b\x04\x59\xfe\x7e\x6e\x0b"
"\x09\x03\x39\xec\x0a\xa6\xfa\xef"
"\xd5\xcc\xc2\xc6\xf4\xce\x8e\x94"
"\x1e\x36\xb2\x6b\xd1\xeb\xc6\x70"
"\xd1\xbd\x1d\x66\x56\x20\xab\xf7"
"\x4f\x78\xa7\xf6\xd2\x98\x09\x58"
"\x5a\x97\xda\xec\x58\xc6\xb0\x50",
.len = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x60\x3d\xeb\x10\x15\xca\x71\xbe"
"\x2b\x73\xae\xf0\x85\x7d\x77\x81"
"\x1f\x35\x2c\x07\x3b\x61\x08\xd7"
"\x2d\x98\x10\xa3\x09\x14\xdf\xf4",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\x60\x1e\xc3\x13\x77\x57\x89\xa5"
"\xb7\xa7\xf5\x04\xbb\xf3\xd2\x28"
"\xf4\x43\xe3\xca\x4d\x62\xb5\x9a"
"\xca\x84\xe9\x90\xca\xca\xf5\xc5"
"\x2b\x09\x30\xda\xa2\x3d\xe9\x4c"
"\xe8\x70\x17\xba\x2d\x84\x98\x8d"
"\xdf\xc9\xc5\x8d\xb6\x7a\xad\xa6"
"\x13\xc2\xdd\x08\x45\x79\x41\xa6",
.len = 64,
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
}, { /* Generated with Crypto++ */
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55"
"\x0F\x32\x55\x78\x9B\xBE\x78\x9B"
"\xBE\xE1\x04\x27\xE1\x04\x27\x4A"
"\x6D\x90\x4A\x6D\x90\xB3\xD6\xF9",
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB"
"\x54\xE0\x49\xB2\x1B\xA7\x10\x79"
"\x05\x6E\xD7\x40\xCC\x35\x9E\x07"
"\x93\xFC\x65\xF1\x5A\xC3\x2C\xB8"
"\x21\x8A\x16\x7F\xE8\x51\xDD\x46"
"\xAF\x18\xA4\x0D\x76\x02\x6B\xD4"
"\x3D\xC9\x32\x9B\x04\x90\xF9\x62"
"\xEE\x57\xC0\x29\xB5\x1E\x87\x13"
"\x7C\xE5\x4E\xDA\x43\xAC\x15\xA1"
"\x0A\x73\xFF\x68\xD1\x3A\xC6\x2F"
"\x98\x01\x8D\xF6\x5F\xEB\x54\xBD"
"\x26\xB2\x1B\x84\x10\x79\xE2\x4B"
"\xD7\x40\xA9\x12\x9E\x07\x70\xFC"
"\x65\xCE\x37\xC3\x2C\x95\x21\x8A"
"\xF3\x5C\xE8\x51\xBA\x23\xAF\x18"
"\x81\x0D\x76\xDF\x48\xD4\x3D\xA6"
"\x0F\x9B\x04\x6D\xF9\x62\xCB\x34"
"\xC0\x29\x92\x1E\x87\xF0\x59\xE5"
"\x4E\xB7\x20\xAC\x15\x7E\x0A\x73"
"\xDC\x45\xD1\x3A\xA3\x0C\x98\x01"
"\x6A\xF6\x5F\xC8\x31\xBD\x26\x8F"
"\x1B\x84\xED\x56\xE2\x4B\xB4\x1D"
"\xA9\x12\x7B\x07\x70\xD9\x42\xCE"
"\x37\xA0\x09\x95\xFE\x67\xF3\x5C"
"\xC5\x2E\xBA\x23\x8C\x18\x81\xEA"
"\x53\xDF\x48\xB1\x1A\xA6\x0F\x78"
"\x04\x6D\xD6\x3F\xCB\x34\x9D\x06"
"\x92\xFB\x64\xF0\x59\xC2\x2B\xB7"
"\x20\x89\x15\x7E\xE7\x50\xDC\x45"
"\xAE\x17\xA3\x0C\x75\x01\x6A\xD3"
"\x3C\xC8\x31\x9A\x03\x8F\xF8\x61"
"\xED\x56\xBF\x28\xB4\x1D\x86\x12",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x04\xF3\xD3\x88\x17\xEF\xDC\xEF"
"\x8B\x04\xF8\x3A\x66\x8D\x1A\x53"
"\x57\x1F\x4B\x23\xE4\xA0\xAF\xF9"
"\x69\x95\x35\x98\x8D\x4D\x8C\xC1"
"\xF0\xB2\x7F\x80\xBB\x54\x28\xA2"
"\x7A\x1B\x9F\x77\xEC\x0E\x6E\xDE"
"\xF0\xEC\xB8\xE4\x20\x62\xEE\xDB"
"\x5D\xF5\xDD\xE3\x54\xFC\xDD\xEB"
"\x6A\xEE\x65\xA1\x21\xD6\xD7\x81"
"\x47\x61\x12\x4D\xC2\x8C\xFA\x78"
"\x1F\x28\x02\x01\xC3\xFC\x1F\xEC"
"\x0F\x10\x4F\xB3\x12\x45\xC6\x3B"
"\x7E\x08\xF9\x5A\xD0\x5D\x73\x2D"
"\x58\xA4\xE5\xCB\x1C\xB4\xCE\x74"
"\x32\x41\x1F\x31\x9C\x08\xA2\x5D"
"\x67\xEB\x72\x1D\xF8\xE7\x70\x54"
"\x34\x4B\x31\x69\x84\x66\x96\x44"
"\x56\xCC\x1E\xD9\xE6\x13\x6A\xB9"
"\x2D\x0A\x05\x45\x2D\x90\xCC\xDF"
"\x16\x5C\x5F\x79\x34\x52\x54\xFE"
"\xFE\xCD\xAD\x04\x2E\xAD\x86\x06"
"\x1F\x37\xE8\x28\xBC\xD3\x8F\x5B"
"\x92\x66\x87\x3B\x8A\x0A\x1A\xCC"
"\x6E\xAB\x9F\x0B\xFA\x5C\xE6\xFD"
"\x3C\x98\x08\x12\xEC\xAA\x9E\x11"
"\xCA\xB2\x1F\xCE\x5E\x5B\xB2\x72"
"\x9C\xCC\x5D\xC5\xE0\x32\xC0\x56"
"\xD5\x45\x16\xD2\xAF\x13\x66\xF7"
"\x8C\x67\xAC\x79\xB2\xAF\x56\x27"
"\x3F\xCC\xFE\xCB\x1E\xC0\x75\xF1"
"\xA7\xC9\xC3\x1D\x8E\xDD\xF9\xD4"
"\x42\xC8\x21\x08\x16\xF7\x01\xD7"
"\xAC\x8E\x3F\x1D\x56\xC1\x06\xE4"
"\x9C\x62\xD6\xA5\x6A\x50\x44\xB3"
"\x35\x1C\x82\xB9\x10\xF9\x42\xA1"
"\xFC\x74\x9B\x44\x4F\x25\x02\xE3"
"\x08\xF5\xD4\x32\x39\x08\x11\xE8"
"\xD2\x6B\x50\x53\xD4\x08\xD1\x6B"
"\x3A\x4A\x68\x7B\x7C\xCD\x46\x5E"
"\x0D\x07\x19\xDB\x67\xD7\x98\x91"
"\xD7\x17\x10\x9B\x7B\x8A\x9B\x33"
"\xAE\xF3\x00\xA6\xD4\x15\xD9\xEA"
"\x85\x99\x22\xE8\x91\x38\x70\x83"
"\x93\x01\x24\x6C\xFA\x9A\xB9\x07"
"\xEA\x8D\x3B\xD9\x2A\x43\x59\x16"
"\x2F\x69\xEE\x84\x36\x44\x76\x98"
"\xF3\x04\x2A\x7C\x74\x3D\x29\x2B"
"\x0D\xAD\x8F\x44\x82\x9E\x57\x8D"
"\xAC\xED\x18\x1F\x50\xA4\xF5\x98"
"\x1F\xBD\x92\x91\x1B\x2D\xA6\xD6"
"\xD2\xE3\x02\xAA\x92\x3B\xC6\xB3"
"\x1B\x39\x72\xD5\x26\xCA\x04\xE0"
"\xFC\x58\x78\xBB\xB1\x3F\xA1\x9C"
"\x42\x24\x3E\x2E\x22\xBB\x4B\xBA"
"\xF4\x52\x0A\xE6\xAE\x47\xB4\x7D"
"\x1D\xA8\xBE\x81\x1A\x75\xDA\xAC"
"\xA6\x25\x1E\xEF\x3A\xC0\x6C\x63"
"\xEF\xDC\xC9\x79\x10\x26\xE8\x61"
"\x29\xFC\xA4\x05\xDF\x7D\x5C\x63"
"\x10\x09\x9B\x46\x9B\xF2\x2C\x2B"
"\xFA\x3A\x05\x4C\xFA\xD1\xFF\xFE"
"\xF1\x4C\xE5\xB2\x91\x64\x0C\x51",
.len = 496,
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
}, { /* Generated with Crypto++ */
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xC9\x83\xA6\xC9\xEC\x0F\x32\x55"
"\x0F\x32\x55\x78\x9B\xBE\x78\x9B"
"\xBE\xE1\x04\x27\xE1\x04\x27\x4A"
"\x6D\x90\x4A\x6D\x90\xB3\xD6\xF9",
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE7\x82\x1D\xB8\x53\x11\xAC\x47"
"\xE2\x7D\x18\xD6\x71\x0C\xA7\x42",
.ptext = "\x50\xB9\x22\xAE\x17\x80\x0C\x75"
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
"\xDE\x47\xD3\x3C\xA5\x0E\x9A\x03"
"\x6C\xF8\x61\xCA\x33\xBF\x28\x91"
"\x1D\x86\xEF\x58\xE4\x4D\xB6\x1F"
"\xAB\x14\x7D\x09\x72\xDB\x44\xD0"
"\x39\xA2\x0B\x97\x00\x69\xF5\x5E"
"\xC7\x30\xBC\x25\x8E\x1A\x83\xEC"
"\x55\xE1\x4A\xB3\x1C\xA8\x11\x7A"
"\x06\x6F\xD8\x41\xCD\x36\x9F\x08"
"\x94\xFD\x66\xF2\x5B\xC4\x2D\xB9"
"\x22\x8B\x17\x80\xE9\x52\xDE\x47"
"\xB0\x19\xA5\x0E\x77\x03\x6C\xD5"
"\x3E\xCA\x33\x9C\x05\x91\xFA\x63"
"\xEF\x58\xC1\x2A\xB6\x1F\x88\x14"
"\x7D\xE6\x4F\xDB\x44\xAD\x16\xA2"
"\x0B\x74\x00\x69\xD2\x3B\xC7\x30"
"\x99\x02\x8E\xF7\x60\xEC\x55\xBE"
"\x27\xB3\x1C\x85\x11\x7A\xE3\x4C"
"\xD8\x41\xAA\x13\x9F\x08\x71\xFD"
"\x66\xCF\x38\xC4\x2D\x96\x22\x8B"
"\xF4\x5D\xE9\x52\xBB\x24\xB0\x19"
"\x82\x0E\x77\xE0\x49\xD5\x3E\xA7"
"\x10\x9C\x05\x6E\xFA\x63\xCC\x35"
"\xC1\x2A\x93\x1F\x88\xF1\x5A\xE6"
"\x4F\xB8\x21\xAD\x16\x7F\x0B\x74"
"\xDD\x46\xD2\x3B\xA4\x0D\x99\x02"
"\x6B\xF7\x60\xC9\x32\xBE\x27\x90"
"\x1C\x85\xEE\x57\xE3\x4C\xB5\x1E"
"\xAA\x13\x7C\x08\x71\xDA\x43\xCF"
"\x38\xA1\x0A\x96\xFF\x68\xF4\x5D"
"\xC6\x2F\xBB\x24\x8D\x19\x82\xEB"
"\x54\xE0\x49\xB2\x1B\xA7\x10\x79"
"\x05\x6E\xD7\x40\xCC\x35\x9E\x07"
"\x93\xFC\x65\xF1\x5A\xC3\x2C\xB8"
"\x21\x8A\x16\x7F\xE8\x51\xDD\x46"
"\xAF\x18\xA4\x0D\x76\x02\x6B\xD4"
"\x3D\xC9\x32\x9B\x04\x90\xF9\x62"
"\xEE\x57\xC0\x29\xB5\x1E\x87\x13"
"\x7C\xE5\x4E\xDA\x43\xAC\x15\xA1"
"\x0A\x73\xFF\x68\xD1\x3A\xC6\x2F"
"\x98\x01\x8D\xF6\x5F\xEB\x54\xBD"
"\x26\xB2\x1B\x84\x10\x79\xE2\x4B"
"\xD7\x40\xA9\x12\x9E\x07\x70\xFC"
"\x65\xCE\x37\xC3\x2C\x95\x21\x8A"
"\xF3\x5C\xE8\x51\xBA\x23\xAF\x18"
"\x81\x0D\x76\xDF\x48\xD4\x3D\xA6"
"\x0F\x9B\x04\x6D\xF9\x62\xCB\x34"
"\xC0\x29\x92\x1E\x87\xF0\x59\xE5"
"\x4E\xB7\x20\xAC\x15\x7E\x0A\x73"
"\xDC\x45\xD1\x3A\xA3\x0C\x98\x01"
"\x6A\xF6\x5F\xC8\x31\xBD\x26\x8F"
"\x1B\x84\xED\x56\xE2\x4B\xB4\x1D"
"\xA9\x12\x7B\x07\x70\xD9\x42\xCE"
"\x37\xA0\x09\x95\xFE\x67\xF3\x5C"
"\xC5\x2E\xBA\x23\x8C\x18\x81\xEA"
"\x53\xDF\x48\xB1\x1A\xA6\x0F\x78"
"\x04\x6D\xD6\x3F\xCB\x34\x9D\x06"
"\x92\xFB\x64\xF0\x59\xC2\x2B\xB7"
"\x20\x89\x15\x7E\xE7\x50\xDC\x45"
"\xAE\x17\xA3\x0C\x75\x01\x6A\xD3"
"\x3C\xC8\x31\x9A\x03\x8F\xF8\x61"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xED\x56\xBF\x28\xB4\x1D\x86\x12"
"\x7B\xE4\x4D",
.ctext = "\xDA\x4E\x3F\xBC\xE8\xB6\x3A\xA2"
"\xD5\x4D\x84\x4A\xA9\x0C\xE1\xA5"
"\xB8\x73\xBC\xF9\xBB\x59\x2F\x44"
"\x8B\xAB\x82\x6C\xB4\x32\x9A\xDE"
"\x5A\x0B\xDB\x7A\x6B\xF2\x38\x9F"
"\x06\xF7\xF7\xFF\xFF\xC0\x8A\x2E"
"\x76\xEA\x06\x32\x23\xF3\x59\x2E"
"\x75\xDE\x71\x86\x3C\x98\x23\x44"
"\x5B\xF2\xFA\x6A\x00\xBB\xC1\xAD"
"\x58\xBD\x3E\x6F\x2E\xB4\x19\x04"
"\x70\x8B\x92\x55\x23\xE9\x6A\x3A"
"\x78\x7A\x1B\x10\x85\x52\x9C\x12"
"\xE4\x55\x81\x21\xCE\x53\xD0\x3B"
"\x63\x77\x2C\x74\xD1\xF5\x60\xF3"
"\xA1\xDE\x44\x3C\x8F\x4D\x2F\xDD"
"\x8A\xFE\x3C\x42\x8E\xD3\xF2\x8E"
"\xA8\x28\x69\x65\x31\xE1\x45\x83"
"\xE4\x49\xC4\x9C\xA7\x28\xAA\x21"
"\xCD\x5D\x0F\x15\xB7\x93\x07\x26"
"\xB0\x65\x6D\x91\x90\x23\x7A\xC6"
"\xDB\x68\xB0\xA1\x8E\xA4\x76\x4E"
"\xC6\x91\x83\x20\x92\x4D\x63\x7A"
"\x45\x18\x18\x74\x19\xAD\x71\x01"
"\x6B\x23\xAD\x9D\x4E\xE4\x6E\x46"
"\xC9\x73\x7A\xF9\x02\x95\xF4\x07"
"\x0E\x7A\xA6\xC5\xAE\xFA\x15\x2C"
"\x51\x71\xF1\xDC\x22\xB6\xAC\xD8"
"\x19\x24\x44\xBC\x0C\xFB\x3C\x2D"
"\xB1\x50\x47\x15\x0E\xDB\xB6\xD7"
"\xE8\x61\xE5\x95\x52\x1E\x3E\x49"
"\x70\xE9\x66\x04\x4C\xE1\xAF\xBD"
"\xDD\x15\x3B\x20\x59\x24\xFF\xB0"
"\x39\xAA\xE7\xBF\x23\xA3\x6E\xD5"
"\x15\xF0\x61\x4F\xAE\x89\x10\x58"
"\x5A\x33\x95\x52\x2A\xB5\x77\x9C"
"\xA5\x43\x80\x40\x27\x2D\xAE\xD9"
"\x3F\xE0\x80\x94\x78\x79\xCB\x7E"
"\xAD\x12\x44\x4C\xEC\x27\xB0\xEE"
"\x0B\x05\x2A\x82\x99\x58\xBB\x7A"
"\x8D\x6D\x9D\x8E\xE2\x8E\xE7\x93"
"\x2F\xB3\x09\x8D\x06\xD5\xEE\x70"
"\x16\xAE\x35\xC5\x52\x0F\x46\x1F"
"\x71\xF9\x5E\xF2\x67\xDC\x98\x2F"
"\xA3\x23\xAA\xD5\xD0\x49\xF4\xA6"
"\xF6\xB8\x32\xCD\xD6\x85\x73\x60"
"\x59\x20\xE7\x55\x0E\x91\xE2\x0C"
"\x3F\x1C\xEB\x3D\xDF\x52\x64\xF2"
"\x7D\x8B\x5D\x63\x16\xB9\xB2\x5D"
"\x5E\xAB\xB2\x97\xAB\x78\x44\xE7"
"\xC6\x72\x20\xC5\x90\x9B\xDC\x5D"
"\xB0\xEF\x44\xEF\x87\x31\x8D\xF4"
"\xFB\x81\x5D\xF7\x96\x96\xD4\x50"
"\x89\xA7\xF6\xB9\x67\x76\x40\x9E"
"\x9D\x40\xD5\x2C\x30\xB8\x01\x8F"
"\xE4\x7B\x71\x48\xA9\xA0\xA0\x1D"
"\x87\x52\xA4\x91\xA9\xD7\xA9\x51"
"\xD9\x59\xF7\xCC\x63\x22\xC1\x8D"
"\x84\x7B\xD8\x22\x32\x5C\x6F\x1D"
"\x6E\x9F\xFA\xDD\x49\x40\xDC\x37"
"\x14\x8C\xE1\x80\x1B\xDD\x36\x2A"
"\xD0\xE9\x54\x99\x5D\xBA\x3B\x11"
"\xD8\xFE\xC9\x5B\x5C\x25\xE5\x76"
"\xFB\xF2\x3F",
.len = 499,
crypto: testmgr - add new larger AES testvectors Most AES testvectors are short and do not test parallelised codepaths well. Add larger testvectors to test large crypto operations and to test multi-page crypto with AES. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:53:02 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 2,
.tap = { 499 - 16, 16 },
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec aes_ctr_rfc3686_tv_template[] = {
{ /* From RFC 3686 */
.key = "\xae\x68\x52\xf8\x12\x10\x67\xcc"
"\x4b\xf7\xa5\x76\x55\x77\xf3\x9e"
"\x00\x00\x00\x30",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "Single block msg",
.ctext = "\xe4\x09\x5d\x4f\xb7\xa7\xb3\x79"
"\x2d\x61\x75\xa3\x26\x13\x11\xb8",
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7e\x24\x06\x78\x17\xfa\xe0\xd7"
"\x43\xd6\xce\x1f\x32\x53\x91\x63"
"\x00\x6c\xb6\xdb",
.klen = 20,
.iv = "\xc0\x54\x3b\x59\xda\x48\xd9\x0b",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x51\x04\xa1\x06\x16\x8a\x72\xd9"
"\x79\x0d\x41\xee\x8e\xda\xd3\x88"
"\xeb\x2e\x1e\xfc\x46\xda\x57\xc8"
"\xfc\xe6\x30\xdf\x91\x41\xbe\x28",
.len = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x16\xaf\x5b\x14\x5f\xc9\xf5\x79"
"\xc1\x75\xf9\x3e\x3b\xfb\x0e\xed"
"\x86\x3d\x06\xcc\xfd\xb7\x85\x15"
"\x00\x00\x00\x48",
.klen = 28,
.iv = "\x36\x73\x3c\x14\x7d\x6d\x93\xcb",
.ptext = "Single block msg",
.ctext = "\x4b\x55\x38\x4f\xe2\x59\xc9\xc8"
"\x4e\x79\x35\xa0\x03\xcb\xe9\x28",
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7c\x5c\xb2\x40\x1b\x3d\xc3\x3c"
"\x19\xe7\x34\x08\x19\xe0\xf6\x9c"
"\x67\x8c\x3d\xb8\xe6\xf6\xa9\x1a"
"\x00\x96\xb0\x3b",
.klen = 28,
.iv = "\x02\x0c\x6e\xad\xc2\xcb\x50\x0d",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x45\x32\x43\xfc\x60\x9b\x23\x32"
"\x7e\xdf\xaa\xfa\x71\x31\xcd\x9f"
"\x84\x90\x70\x1c\x5a\xd4\xa7\x9c"
"\xfc\x1f\xe0\xff\x42\xf4\xfb\x00",
.len = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x77\x6b\xef\xf2\x85\x1d\xb0\x6f"
"\x4c\x8a\x05\x42\xc8\x69\x6f\x6c"
"\x6a\x81\xaf\x1e\xec\x96\xb4\xd3"
"\x7f\xc1\xd6\x89\xe6\xc1\xc1\x04"
"\x00\x00\x00\x60",
.klen = 36,
.iv = "\xdb\x56\x72\xc9\x7a\xa8\xf0\xb2",
.ptext = "Single block msg",
.ctext = "\x14\x5a\xd0\x1d\xbf\x82\x4e\xc7"
"\x56\x08\x63\xdc\x71\xe3\xe0\xc0",
.len = 16,
crypto: testmgr - add aead null encryption test vectors Add test vectors for aead with null encryption and md5, respectively sha1 authentication. Input data is taken from test vectors listed in RFC2410. Signed-off-by: Horia Geanta <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-03-14 09:46:51 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf6\xd6\x6d\x6b\xd5\x2d\x59\xbb"
"\x07\x96\x36\x58\x79\xef\xf8\x86"
"\xc6\x6d\xd5\x1a\x5b\x6a\x99\x74"
"\x4b\x50\x59\x0c\x87\xa2\x38\x84"
"\x00\xfa\xac\x24",
.klen = 36,
.iv = "\xc1\x58\x5e\xf1\x5a\x43\xd8\x75",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add aead cbc aes hmac sha1,256,512 test vectors Test vectors were generated starting from existing CBC(AES) test vectors (RFC3602, NIST SP800-38A) and adding HMAC(SHA*) computed with Crypto++ and double-checked with HashCalc. Signed-off-by: Horia Geanta <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-03 10:16:54 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xf0\x5e\x23\x1b\x38\x94\x61\x2c"
"\x49\xee\x00\x0b\x80\x4e\xb2\xa9"
"\xb8\x30\x6b\x50\x8f\x83\x9d\x6a"
"\x55\x30\x83\x1d\x93\x44\xaf\x1c",
.len = 32,
crypto: testmgr - add aead null encryption test vectors Add test vectors for aead with null encryption and md5, respectively sha1 authentication. Input data is taken from test vectors listed in RFC2410. Signed-off-by: Horia Geanta <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-03-14 09:46:51 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
// generated using Crypto++
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x00\x00\x00\x00",
.klen = 32 + 4,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext =
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x03\x06\x09\x0c\x0f\x12\x15"
"\x18\x1b\x1e\x21\x24\x27\x2a\x2d"
"\x30\x33\x36\x39\x3c\x3f\x42\x45"
"\x48\x4b\x4e\x51\x54\x57\x5a\x5d"
"\x60\x63\x66\x69\x6c\x6f\x72\x75"
"\x78\x7b\x7e\x81\x84\x87\x8a\x8d"
"\x90\x93\x96\x99\x9c\x9f\xa2\xa5"
"\xa8\xab\xae\xb1\xb4\xb7\xba\xbd"
"\xc0\xc3\xc6\xc9\xcc\xcf\xd2\xd5"
"\xd8\xdb\xde\xe1\xe4\xe7\xea\xed"
"\xf0\xf3\xf6\xf9\xfc\xff\x02\x05"
"\x08\x0b\x0e\x11\x14\x17\x1a\x1d"
"\x20\x23\x26\x29\x2c\x2f\x32\x35"
"\x38\x3b\x3e\x41\x44\x47\x4a\x4d"
"\x50\x53\x56\x59\x5c\x5f\x62\x65"
"\x68\x6b\x6e\x71\x74\x77\x7a\x7d"
"\x80\x83\x86\x89\x8c\x8f\x92\x95"
"\x98\x9b\x9e\xa1\xa4\xa7\xaa\xad"
"\xb0\xb3\xb6\xb9\xbc\xbf\xc2\xc5"
"\xc8\xcb\xce\xd1\xd4\xd7\xda\xdd"
"\xe0\xe3\xe6\xe9\xec\xef\xf2\xf5"
"\xf8\xfb\xfe\x01\x04\x07\x0a\x0d"
"\x10\x13\x16\x19\x1c\x1f\x22\x25"
"\x28\x2b\x2e\x31\x34\x37\x3a\x3d"
"\x40\x43\x46\x49\x4c\x4f\x52\x55"
"\x58\x5b\x5e\x61\x64\x67\x6a\x6d"
"\x70\x73\x76\x79\x7c\x7f\x82\x85"
"\x88\x8b\x8e\x91\x94\x97\x9a\x9d"
"\xa0\xa3\xa6\xa9\xac\xaf\xb2\xb5"
"\xb8\xbb\xbe\xc1\xc4\xc7\xca\xcd"
"\xd0\xd3\xd6\xd9\xdc\xdf\xe2\xe5"
"\xe8\xeb\xee\xf1\xf4\xf7\xfa\xfd"
"\x00\x05\x0a\x0f\x14\x19\x1e\x23"
"\x28\x2d\x32\x37\x3c\x41\x46\x4b"
"\x50\x55\x5a\x5f\x64\x69\x6e\x73"
"\x78\x7d\x82\x87\x8c\x91\x96\x9b"
"\xa0\xa5\xaa\xaf\xb4\xb9\xbe\xc3"
"\xc8\xcd\xd2\xd7\xdc\xe1\xe6\xeb"
"\xf0\xf5\xfa\xff\x04\x09\x0e\x13"
"\x18\x1d\x22\x27\x2c\x31\x36\x3b"
"\x40\x45\x4a\x4f\x54\x59\x5e\x63"
"\x68\x6d\x72\x77\x7c\x81\x86\x8b"
"\x90\x95\x9a\x9f\xa4\xa9\xae\xb3"
"\xb8\xbd\xc2\xc7\xcc\xd1\xd6\xdb"
"\xe0\xe5\xea\xef\xf4\xf9\xfe\x03"
"\x08\x0d\x12\x17\x1c\x21\x26\x2b"
"\x30\x35\x3a\x3f\x44\x49\x4e\x53"
"\x58\x5d\x62\x67\x6c\x71\x76\x7b"
"\x80\x85\x8a\x8f\x94\x99\x9e\xa3"
"\xa8\xad\xb2\xb7\xbc\xc1\xc6\xcb"
"\xd0\xd5\xda\xdf\xe4\xe9\xee\xf3"
"\xf8\xfd\x02\x07\x0c\x11\x16\x1b"
"\x20\x25\x2a\x2f\x34\x39\x3e\x43"
"\x48\x4d\x52\x57\x5c\x61\x66\x6b"
"\x70\x75\x7a\x7f\x84\x89\x8e\x93"
"\x98\x9d\xa2\xa7\xac\xb1\xb6\xbb"
"\xc0\xc5\xca\xcf\xd4\xd9\xde\xe3"
"\xe8\xed\xf2\xf7\xfc\x01\x06\x0b"
"\x10\x15\x1a\x1f\x24\x29\x2e\x33"
"\x38\x3d\x42\x47\x4c\x51\x56\x5b"
"\x60\x65\x6a\x6f\x74\x79\x7e\x83"
"\x88\x8d\x92\x97\x9c\xa1\xa6\xab"
"\xb0\xb5\xba\xbf\xc4\xc9\xce\xd3"
"\xd8\xdd\xe2\xe7\xec\xf1\xf6\xfb"
"\x00\x07\x0e\x15\x1c\x23\x2a\x31"
"\x38\x3f\x46\x4d\x54\x5b\x62\x69"
"\x70\x77\x7e\x85\x8c\x93\x9a\xa1"
"\xa8\xaf\xb6\xbd\xc4\xcb\xd2\xd9"
"\xe0\xe7\xee\xf5\xfc\x03\x0a\x11"
"\x18\x1f\x26\x2d\x34\x3b\x42\x49"
"\x50\x57\x5e\x65\x6c\x73\x7a\x81"
"\x88\x8f\x96\x9d\xa4\xab\xb2\xb9"
"\xc0\xc7\xce\xd5\xdc\xe3\xea\xf1"
"\xf8\xff\x06\x0d\x14\x1b\x22\x29"
"\x30\x37\x3e\x45\x4c\x53\x5a\x61"
"\x68\x6f\x76\x7d\x84\x8b\x92\x99"
"\xa0\xa7\xae\xb5\xbc\xc3\xca\xd1"
"\xd8\xdf\xe6\xed\xf4\xfb\x02\x09"
"\x10\x17\x1e\x25\x2c\x33\x3a\x41"
"\x48\x4f\x56\x5d\x64\x6b\x72\x79"
"\x80\x87\x8e\x95\x9c\xa3\xaa\xb1"
"\xb8\xbf\xc6\xcd\xd4\xdb\xe2\xe9"
"\xf0\xf7\xfe\x05\x0c\x13\x1a\x21"
"\x28\x2f\x36\x3d\x44\x4b\x52\x59"
"\x60\x67\x6e\x75\x7c\x83\x8a\x91"
"\x98\x9f\xa6\xad\xb4\xbb\xc2\xc9"
"\xd0\xd7\xde\xe5\xec\xf3\xfa\x01"
"\x08\x0f\x16\x1d\x24\x2b\x32\x39"
"\x40\x47\x4e\x55\x5c\x63\x6a\x71"
"\x78\x7f\x86\x8d\x94\x9b\xa2\xa9"
"\xb0\xb7\xbe\xc5\xcc\xd3\xda\xe1"
"\xe8\xef\xf6\xfd\x04\x0b\x12\x19"
"\x20\x27\x2e\x35\x3c\x43\x4a\x51"
"\x58\x5f\x66\x6d\x74\x7b\x82\x89"
"\x90\x97\x9e\xa5\xac\xb3\xba\xc1"
"\xc8\xcf\xd6\xdd\xe4\xeb\xf2\xf9"
"\x00\x09\x12\x1b\x24\x2d\x36\x3f"
"\x48\x51\x5a\x63\x6c\x75\x7e\x87"
"\x90\x99\xa2\xab\xb4\xbd\xc6\xcf"
"\xd8\xe1\xea\xf3\xfc\x05\x0e\x17"
"\x20\x29\x32\x3b\x44\x4d\x56\x5f"
"\x68\x71\x7a\x83\x8c\x95\x9e\xa7"
"\xb0\xb9\xc2\xcb\xd4\xdd\xe6\xef"
"\xf8\x01\x0a\x13\x1c\x25\x2e\x37"
"\x40\x49\x52\x5b\x64\x6d\x76\x7f"
"\x88\x91\x9a\xa3\xac\xb5\xbe\xc7"
"\xd0\xd9\xe2\xeb\xf4\xfd\x06\x0f"
"\x18\x21\x2a\x33\x3c\x45\x4e\x57"
"\x60\x69\x72\x7b\x84\x8d\x96\x9f"
"\xa8\xb1\xba\xc3\xcc\xd5\xde\xe7"
"\xf0\xf9\x02\x0b\x14\x1d\x26\x2f"
"\x38\x41\x4a\x53\x5c\x65\x6e\x77"
"\x80\x89\x92\x9b\xa4\xad\xb6\xbf"
"\xc8\xd1\xda\xe3\xec\xf5\xfe\x07"
"\x10\x19\x22\x2b\x34\x3d\x46\x4f"
"\x58\x61\x6a\x73\x7c\x85\x8e\x97"
"\xa0\xa9\xb2\xbb\xc4\xcd\xd6\xdf"
"\xe8\xf1\xfa\x03\x0c\x15\x1e\x27"
"\x30\x39\x42\x4b\x54\x5d\x66\x6f"
"\x78\x81\x8a\x93\x9c\xa5\xae\xb7"
"\xc0\xc9\xd2\xdb\xe4\xed\xf6\xff"
"\x08\x11\x1a\x23\x2c\x35\x3e\x47"
"\x50\x59\x62\x6b\x74\x7d\x86\x8f"
"\x98\xa1\xaa\xb3\xbc\xc5\xce\xd7"
"\xe0\xe9\xf2\xfb\x04\x0d\x16\x1f"
"\x28\x31\x3a\x43\x4c\x55\x5e\x67"
"\x70\x79\x82\x8b\x94\x9d\xa6\xaf"
"\xb8\xc1\xca\xd3\xdc\xe5\xee\xf7"
"\x00\x0b\x16\x21\x2c\x37\x42\x4d"
"\x58\x63\x6e\x79\x84\x8f\x9a\xa5"
"\xb0\xbb\xc6\xd1\xdc\xe7\xf2\xfd"
"\x08\x13\x1e\x29\x34\x3f\x4a\x55"
"\x60\x6b\x76\x81\x8c\x97\xa2\xad"
"\xb8\xc3\xce\xd9\xe4\xef\xfa\x05"
"\x10\x1b\x26\x31\x3c\x47\x52\x5d"
"\x68\x73\x7e\x89\x94\x9f\xaa\xb5"
"\xc0\xcb\xd6\xe1\xec\xf7\x02\x0d"
"\x18\x23\x2e\x39\x44\x4f\x5a\x65"
"\x70\x7b\x86\x91\x9c\xa7\xb2\xbd"
"\xc8\xd3\xde\xe9\xf4\xff\x0a\x15"
"\x20\x2b\x36\x41\x4c\x57\x62\x6d"
"\x78\x83\x8e\x99\xa4\xaf\xba\xc5"
"\xd0\xdb\xe6\xf1\xfc\x07\x12\x1d"
"\x28\x33\x3e\x49\x54\x5f\x6a\x75"
"\x80\x8b\x96\xa1\xac\xb7\xc2\xcd"
"\xd8\xe3\xee\xf9\x04\x0f\x1a\x25"
"\x30\x3b\x46\x51\x5c\x67\x72\x7d"
"\x88\x93\x9e\xa9\xb4\xbf\xca\xd5"
"\xe0\xeb\xf6\x01\x0c\x17\x22\x2d"
"\x38\x43\x4e\x59\x64\x6f\x7a\x85"
"\x90\x9b\xa6\xb1\xbc\xc7\xd2\xdd"
"\xe8\xf3\xfe\x09\x14\x1f\x2a\x35"
"\x40\x4b\x56\x61\x6c\x77\x82\x8d"
"\x98\xa3\xae\xb9\xc4\xcf\xda\xe5"
"\xf0\xfb\x06\x11\x1c\x27\x32\x3d"
"\x48\x53\x5e\x69\x74\x7f\x8a\x95"
"\xa0\xab\xb6\xc1\xcc\xd7\xe2\xed"
"\xf8\x03\x0e\x19\x24\x2f\x3a\x45"
"\x50\x5b\x66\x71\x7c\x87\x92\x9d"
"\xa8\xb3\xbe\xc9\xd4\xdf\xea\xf5"
"\x00\x0d\x1a\x27\x34\x41\x4e\x5b"
"\x68\x75\x82\x8f\x9c\xa9\xb6\xc3"
"\xd0\xdd\xea\xf7\x04\x11\x1e\x2b"
"\x38\x45\x52\x5f\x6c\x79\x86\x93"
"\xa0\xad\xba\xc7\xd4\xe1\xee\xfb"
"\x08\x15\x22\x2f\x3c\x49\x56\x63"
"\x70\x7d\x8a\x97\xa4\xb1\xbe\xcb"
"\xd8\xe5\xf2\xff\x0c\x19\x26\x33"
"\x40\x4d\x5a\x67\x74\x81\x8e\x9b"
"\xa8\xb5\xc2\xcf\xdc\xe9\xf6\x03"
"\x10\x1d\x2a\x37\x44\x51\x5e\x6b"
"\x78\x85\x92\x9f\xac\xb9\xc6\xd3"
"\xe0\xed\xfa\x07\x14\x21\x2e\x3b"
"\x48\x55\x62\x6f\x7c\x89\x96\xa3"
"\xb0\xbd\xca\xd7\xe4\xf1\xfe\x0b"
"\x18\x25\x32\x3f\x4c\x59\x66\x73"
"\x80\x8d\x9a\xa7\xb4\xc1\xce\xdb"
"\xe8\xf5\x02\x0f\x1c\x29\x36\x43"
"\x50\x5d\x6a\x77\x84\x91\x9e\xab"
"\xb8\xc5\xd2\xdf\xec\xf9\x06\x13"
"\x20\x2d\x3a\x47\x54\x61\x6e\x7b"
"\x88\x95\xa2\xaf\xbc\xc9\xd6\xe3"
"\xf0\xfd\x0a\x17\x24\x31\x3e\x4b"
"\x58\x65\x72\x7f\x8c\x99\xa6\xb3"
"\xc0\xcd\xda\xe7\xf4\x01\x0e\x1b"
"\x28\x35\x42\x4f\x5c\x69\x76\x83"
"\x90\x9d\xaa\xb7\xc4\xd1\xde\xeb"
"\xf8\x05\x12\x1f\x2c\x39\x46\x53"
"\x60\x6d\x7a\x87\x94\xa1\xae\xbb"
"\xc8\xd5\xe2\xef\xfc\x09\x16\x23"
"\x30\x3d\x4a\x57\x64\x71\x7e\x8b"
"\x98\xa5\xb2\xbf\xcc\xd9\xe6\xf3"
"\x00\x0f\x1e\x2d\x3c\x4b\x5a\x69"
"\x78\x87\x96\xa5\xb4\xc3\xd2\xe1"
"\xf0\xff\x0e\x1d\x2c\x3b\x4a\x59"
"\x68\x77\x86\x95\xa4\xb3\xc2\xd1"
"\xe0\xef\xfe\x0d\x1c\x2b\x3a\x49"
"\x58\x67\x76\x85\x94\xa3\xb2\xc1"
"\xd0\xdf\xee\xfd\x0c\x1b\x2a\x39"
"\x48\x57\x66\x75\x84\x93\xa2\xb1"
"\xc0\xcf\xde\xed\xfc\x0b\x1a\x29"
"\x38\x47\x56\x65\x74\x83\x92\xa1"
"\xb0\xbf\xce\xdd\xec\xfb\x0a\x19"
"\x28\x37\x46\x55\x64\x73\x82\x91"
"\xa0\xaf\xbe\xcd\xdc\xeb\xfa\x09"
"\x18\x27\x36\x45\x54\x63\x72\x81"
"\x90\x9f\xae\xbd\xcc\xdb\xea\xf9"
"\x08\x17\x26\x35\x44\x53\x62\x71"
"\x80\x8f\x9e\xad\xbc\xcb\xda\xe9"
"\xf8\x07\x16\x25\x34\x43\x52\x61"
"\x70\x7f\x8e\x9d\xac\xbb\xca\xd9"
"\xe8\xf7\x06\x15\x24\x33\x42\x51"
"\x60\x6f\x7e\x8d\x9c\xab\xba\xc9"
"\xd8\xe7\xf6\x05\x14\x23\x32\x41"
"\x50\x5f\x6e\x7d\x8c\x9b\xaa\xb9"
"\xc8\xd7\xe6\xf5\x04\x13\x22\x31"
"\x40\x4f\x5e\x6d\x7c\x8b\x9a\xa9"
"\xb8\xc7\xd6\xe5\xf4\x03\x12\x21"
"\x30\x3f\x4e\x5d\x6c\x7b\x8a\x99"
"\xa8\xb7\xc6\xd5\xe4\xf3\x02\x11"
"\x20\x2f\x3e\x4d\x5c\x6b\x7a\x89"
"\x98\xa7\xb6\xc5\xd4\xe3\xf2\x01"
"\x10\x1f\x2e\x3d\x4c\x5b\x6a\x79"
"\x88\x97\xa6\xb5\xc4\xd3\xe2\xf1"
"\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff"
"\x10\x21\x32\x43\x54\x65\x76\x87"
"\x98\xa9\xba\xcb\xdc\xed\xfe\x0f"
"\x20\x31\x42\x53\x64\x75\x86\x97"
"\xa8\xb9\xca\xdb\xec\xfd\x0e\x1f"
"\x30\x41\x52\x63\x74\x85\x96\xa7"
"\xb8\xc9\xda\xeb\xfc\x0d\x1e\x2f"
"\x40\x51\x62\x73\x84\x95\xa6\xb7"
"\xc8\xd9\xea\xfb\x0c\x1d\x2e\x3f"
"\x50\x61\x72\x83\x94\xa5\xb6\xc7"
"\xd8\xe9\xfa\x0b\x1c\x2d\x3e\x4f"
"\x60\x71\x82\x93\xa4\xb5\xc6\xd7"
"\xe8\xf9\x0a\x1b\x2c\x3d\x4e\x5f"
"\x70\x81\x92\xa3\xb4\xc5\xd6\xe7"
"\xf8\x09\x1a\x2b\x3c\x4d\x5e\x6f"
"\x80\x91\xa2\xb3\xc4\xd5\xe6\xf7"
"\x08\x19\x2a\x3b\x4c\x5d\x6e\x7f"
"\x90\xa1\xb2\xc3\xd4\xe5\xf6\x07"
"\x18\x29\x3a\x4b\x5c\x6d\x7e\x8f"
"\xa0\xb1\xc2\xd3\xe4\xf5\x06\x17"
"\x28\x39\x4a\x5b\x6c\x7d\x8e\x9f"
"\xb0\xc1\xd2\xe3\xf4\x05\x16\x27"
"\x38\x49\x5a\x6b\x7c\x8d\x9e\xaf"
"\xc0\xd1\xe2\xf3\x04\x15\x26\x37"
"\x48\x59\x6a\x7b\x8c\x9d\xae\xbf"
"\xd0\xe1\xf2\x03\x14\x25\x36\x47"
"\x58\x69\x7a\x8b\x9c\xad\xbe\xcf"
"\xe0\xf1\x02\x13\x24\x35\x46\x57"
"\x68\x79\x8a\x9b\xac\xbd\xce\xdf"
"\xf0\x01\x12\x23\x34\x45\x56\x67"
"\x78\x89\x9a\xab\xbc\xcd\xde\xef"
"\x00\x13\x26\x39\x4c\x5f\x72\x85"
"\x98\xab\xbe\xd1\xe4\xf7\x0a\x1d"
"\x30\x43\x56\x69\x7c\x8f\xa2\xb5"
"\xc8\xdb\xee\x01\x14\x27\x3a\x4d"
"\x60\x73\x86\x99\xac\xbf\xd2\xe5"
"\xf8\x0b\x1e\x31\x44\x57\x6a\x7d"
"\x90\xa3\xb6\xc9\xdc\xef\x02\x15"
"\x28\x3b\x4e\x61\x74\x87\x9a\xad"
"\xc0\xd3\xe6\xf9\x0c\x1f\x32\x45"
"\x58\x6b\x7e\x91\xa4\xb7\xca\xdd"
"\xf0\x03\x16\x29\x3c\x4f\x62\x75"
"\x88\x9b\xae\xc1\xd4\xe7\xfa\x0d"
"\x20\x33\x46\x59\x6c\x7f\x92\xa5"
"\xb8\xcb\xde\xf1\x04\x17\x2a\x3d"
"\x50\x63\x76\x89\x9c\xaf\xc2\xd5"
"\xe8\xfb\x0e\x21\x34\x47\x5a\x6d"
"\x80\x93\xa6\xb9\xcc\xdf\xf2\x05"
"\x18\x2b\x3e\x51\x64\x77\x8a\x9d"
"\xb0\xc3\xd6\xe9\xfc\x0f\x22\x35"
"\x48\x5b\x6e\x81\x94\xa7\xba\xcd"
"\xe0\xf3\x06\x19\x2c\x3f\x52\x65"
"\x78\x8b\x9e\xb1\xc4\xd7\xea\xfd"
"\x10\x23\x36\x49\x5c\x6f\x82\x95"
"\xa8\xbb\xce\xe1\xf4\x07\x1a\x2d"
"\x40\x53\x66\x79\x8c\x9f\xb2\xc5"
"\xd8\xeb\xfe\x11\x24\x37\x4a\x5d"
"\x70\x83\x96\xa9\xbc\xcf\xe2\xf5"
"\x08\x1b\x2e\x41\x54\x67\x7a\x8d"
"\xa0\xb3\xc6\xd9\xec\xff\x12\x25"
"\x38\x4b\x5e\x71\x84\x97\xaa\xbd"
"\xd0\xe3\xf6\x09\x1c\x2f\x42\x55"
"\x68\x7b\x8e\xa1\xb4\xc7\xda\xed"
"\x00\x15\x2a\x3f\x54\x69\x7e\x93"
"\xa8\xbd\xd2\xe7\xfc\x11\x26\x3b"
"\x50\x65\x7a\x8f\xa4\xb9\xce\xe3"
"\xf8\x0d\x22\x37\x4c\x61\x76\x8b"
"\xa0\xb5\xca\xdf\xf4\x09\x1e\x33"
"\x48\x5d\x72\x87\x9c\xb1\xc6\xdb"
"\xf0\x05\x1a\x2f\x44\x59\x6e\x83"
"\x98\xad\xc2\xd7\xec\x01\x16\x2b"
"\x40\x55\x6a\x7f\x94\xa9\xbe\xd3"
"\xe8\xfd\x12\x27\x3c\x51\x66\x7b"
"\x90\xa5\xba\xcf\xe4\xf9\x0e\x23"
"\x38\x4d\x62\x77\x8c\xa1\xb6\xcb"
"\xe0\xf5\x0a\x1f\x34\x49\x5e\x73"
"\x88\x9d\xb2\xc7\xdc\xf1\x06\x1b"
"\x30\x45\x5a\x6f\x84\x99\xae\xc3"
"\xd8\xed\x02\x17\x2c\x41\x56\x6b"
"\x80\x95\xaa\xbf\xd4\xe9\xfe\x13"
"\x28\x3d\x52\x67\x7c\x91\xa6\xbb"
"\xd0\xe5\xfa\x0f\x24\x39\x4e\x63"
"\x78\x8d\xa2\xb7\xcc\xe1\xf6\x0b"
"\x20\x35\x4a\x5f\x74\x89\x9e\xb3"
"\xc8\xdd\xf2\x07\x1c\x31\x46\x5b"
"\x70\x85\x9a\xaf\xc4\xd9\xee\x03"
"\x18\x2d\x42\x57\x6c\x81\x96\xab"
"\xc0\xd5\xea\xff\x14\x29\x3e\x53"
"\x68\x7d\x92\xa7\xbc\xd1\xe6\xfb"
"\x10\x25\x3a\x4f\x64\x79\x8e\xa3"
"\xb8\xcd\xe2\xf7\x0c\x21\x36\x4b"
"\x60\x75\x8a\x9f\xb4\xc9\xde\xf3"
"\x08\x1d\x32\x47\x5c\x71\x86\x9b"
"\xb0\xc5\xda\xef\x04\x19\x2e\x43"
"\x58\x6d\x82\x97\xac\xc1\xd6\xeb"
"\x00\x17\x2e\x45\x5c\x73\x8a\xa1"
"\xb8\xcf\xe6\xfd\x14\x2b\x42\x59"
"\x70\x87\x9e\xb5\xcc\xe3\xfa\x11"
"\x28\x3f\x56\x6d\x84\x9b\xb2\xc9"
"\xe0\xf7\x0e\x25\x3c\x53\x6a\x81"
"\x98\xaf\xc6\xdd\xf4\x0b\x22\x39"
"\x50\x67\x7e\x95\xac\xc3\xda\xf1"
"\x08\x1f\x36\x4d\x64\x7b\x92\xa9"
"\xc0\xd7\xee\x05\x1c\x33\x4a\x61"
"\x78\x8f\xa6\xbd\xd4\xeb\x02\x19"
"\x30\x47\x5e\x75\x8c\xa3\xba\xd1"
"\xe8\xff\x16\x2d\x44\x5b\x72\x89"
"\xa0\xb7\xce\xe5\xfc\x13\x2a\x41"
"\x58\x6f\x86\x9d\xb4\xcb\xe2\xf9"
"\x10\x27\x3e\x55\x6c\x83\x9a\xb1"
"\xc8\xdf\xf6\x0d\x24\x3b\x52\x69"
"\x80\x97\xae\xc5\xdc\xf3\x0a\x21"
"\x38\x4f\x66\x7d\x94\xab\xc2\xd9"
"\xf0\x07\x1e\x35\x4c\x63\x7a\x91"
"\xa8\xbf\xd6\xed\x04\x1b\x32\x49"
"\x60\x77\x8e\xa5\xbc\xd3\xea\x01"
"\x18\x2f\x46\x5d\x74\x8b\xa2\xb9"
"\xd0\xe7\xfe\x15\x2c\x43\x5a\x71"
"\x88\x9f\xb6\xcd\xe4\xfb\x12\x29"
"\x40\x57\x6e\x85\x9c\xb3\xca\xe1"
"\xf8\x0f\x26\x3d\x54\x6b\x82\x99"
"\xb0\xc7\xde\xf5\x0c\x23\x3a\x51"
"\x68\x7f\x96\xad\xc4\xdb\xf2\x09"
"\x20\x37\x4e\x65\x7c\x93\xaa\xc1"
"\xd8\xef\x06\x1d\x34\x4b\x62\x79"
"\x90\xa7\xbe\xd5\xec\x03\x1a\x31"
"\x48\x5f\x76\x8d\xa4\xbb\xd2\xe9"
"\x00\x19\x32\x4b\x64\x7d\x96\xaf"
"\xc8\xe1\xfa\x13\x2c\x45\x5e\x77"
"\x90\xa9\xc2\xdb\xf4\x0d\x26\x3f"
"\x58\x71\x8a\xa3\xbc\xd5\xee\x07"
"\x20\x39\x52\x6b\x84\x9d\xb6\xcf"
"\xe8\x01\x1a\x33\x4c\x65\x7e\x97"
"\xb0\xc9\xe2\xfb\x14\x2d\x46\x5f"
"\x78\x91\xaa\xc3\xdc\xf5\x0e\x27"
"\x40\x59\x72\x8b\xa4\xbd\xd6\xef"
"\x08\x21\x3a\x53\x6c\x85\x9e\xb7"
"\xd0\xe9\x02\x1b\x34\x4d\x66\x7f"
"\x98\xb1\xca\xe3\xfc\x15\x2e\x47"
"\x60\x79\x92\xab\xc4\xdd\xf6\x0f"
"\x28\x41\x5a\x73\x8c\xa5\xbe\xd7"
"\xf0\x09\x22\x3b\x54\x6d\x86\x9f"
"\xb8\xd1\xea\x03\x1c\x35\x4e\x67"
"\x80\x99\xb2\xcb\xe4\xfd\x16\x2f"
"\x48\x61\x7a\x93\xac\xc5\xde\xf7"
"\x10\x29\x42\x5b\x74\x8d\xa6\xbf"
"\xd8\xf1\x0a\x23\x3c\x55\x6e\x87"
"\xa0\xb9\xd2\xeb\x04\x1d\x36\x4f"
"\x68\x81\x9a\xb3\xcc\xe5\xfe\x17"
"\x30\x49\x62\x7b\x94\xad\xc6\xdf"
"\xf8\x11\x2a\x43\x5c\x75\x8e\xa7"
"\xc0\xd9\xf2\x0b\x24\x3d\x56\x6f"
"\x88\xa1\xba\xd3\xec\x05\x1e\x37"
"\x50\x69\x82\x9b\xb4\xcd\xe6\xff"
"\x18\x31\x4a\x63\x7c\x95\xae\xc7"
"\xe0\xf9\x12\x2b\x44\x5d\x76\x8f"
"\xa8\xc1\xda\xf3\x0c\x25\x3e\x57"
"\x70\x89\xa2\xbb\xd4\xed\x06\x1f"
"\x38\x51\x6a\x83\x9c\xb5\xce\xe7"
"\x00\x1b\x36\x51\x6c\x87\xa2\xbd"
"\xd8\xf3\x0e\x29\x44\x5f\x7a\x95"
"\xb0\xcb\xe6\x01\x1c\x37\x52\x6d"
"\x88\xa3\xbe\xd9\xf4\x0f\x2a\x45"
"\x60\x7b\x96\xb1\xcc\xe7\x02\x1d"
"\x38\x53\x6e\x89\xa4\xbf\xda\xf5"
"\x10\x2b\x46\x61\x7c\x97\xb2\xcd"
"\xe8\x03\x1e\x39\x54\x6f\x8a\xa5"
"\xc0\xdb\xf6\x11\x2c\x47\x62\x7d"
"\x98\xb3\xce\xe9\x04\x1f\x3a\x55"
"\x70\x8b\xa6\xc1\xdc\xf7\x12\x2d"
"\x48\x63\x7e\x99\xb4\xcf\xea\x05"
"\x20\x3b\x56\x71\x8c\xa7\xc2\xdd"
"\xf8\x13\x2e\x49\x64\x7f\x9a\xb5"
"\xd0\xeb\x06\x21\x3c\x57\x72\x8d"
"\xa8\xc3\xde\xf9\x14\x2f\x4a\x65"
"\x80\x9b\xb6\xd1\xec\x07\x22\x3d"
"\x58\x73\x8e\xa9\xc4\xdf\xfa\x15"
"\x30\x4b\x66\x81\x9c\xb7\xd2\xed"
"\x08\x23\x3e\x59\x74\x8f\xaa\xc5"
"\xe0\xfb\x16\x31\x4c\x67\x82\x9d"
"\xb8\xd3\xee\x09\x24\x3f\x5a\x75"
"\x90\xab\xc6\xe1\xfc\x17\x32\x4d"
"\x68\x83\x9e\xb9\xd4\xef\x0a\x25"
"\x40\x5b\x76\x91\xac\xc7\xe2\xfd"
"\x18\x33\x4e\x69\x84\x9f\xba\xd5"
"\xf0\x0b\x26\x41\x5c\x77\x92\xad"
"\xc8\xe3\xfe\x19\x34\x4f\x6a\x85"
"\xa0\xbb\xd6\xf1\x0c\x27\x42\x5d"
"\x78\x93\xae\xc9\xe4\xff\x1a\x35"
"\x50\x6b\x86\xa1\xbc\xd7\xf2\x0d"
"\x28\x43\x5e\x79\x94\xaf\xca\xe5"
"\x00\x1d\x3a\x57\x74\x91\xae\xcb"
"\xe8\x05\x22\x3f\x5c\x79\x96\xb3"
"\xd0\xed\x0a\x27\x44\x61\x7e\x9b"
"\xb8\xd5\xf2\x0f\x2c\x49\x66\x83"
"\xa0\xbd\xda\xf7\x14\x31\x4e\x6b"
"\x88\xa5\xc2\xdf\xfc\x19\x36\x53"
"\x70\x8d\xaa\xc7\xe4\x01\x1e\x3b"
"\x58\x75\x92\xaf\xcc\xe9\x06\x23"
"\x40\x5d\x7a\x97\xb4\xd1\xee\x0b"
"\x28\x45\x62\x7f\x9c\xb9\xd6\xf3"
"\x10\x2d\x4a\x67\x84\xa1\xbe\xdb"
"\xf8\x15\x32\x4f\x6c\x89\xa6\xc3"
"\xe0\xfd\x1a\x37\x54\x71\x8e\xab"
"\xc8\xe5\x02\x1f\x3c\x59\x76\x93"
"\xb0\xcd\xea\x07\x24\x41\x5e\x7b"
"\x98\xb5\xd2\xef\x0c\x29\x46\x63"
"\x80\x9d\xba\xd7\xf4\x11\x2e\x4b"
"\x68\x85\xa2\xbf\xdc\xf9\x16\x33"
"\x50\x6d\x8a\xa7\xc4\xe1\xfe\x1b"
"\x38\x55\x72\x8f\xac\xc9\xe6\x03"
"\x20\x3d\x5a\x77\x94\xb1\xce\xeb"
"\x08\x25\x42\x5f\x7c\x99\xb6\xd3"
"\xf0\x0d\x2a\x47\x64\x81\x9e\xbb"
"\xd8\xf5\x12\x2f\x4c\x69\x86\xa3"
"\xc0\xdd\xfa\x17\x34\x51\x6e\x8b"
"\xa8\xc5\xe2\xff\x1c\x39\x56\x73"
"\x90\xad\xca\xe7\x04\x21\x3e\x5b"
"\x78\x95\xb2\xcf\xec\x09\x26\x43"
"\x60\x7d\x9a\xb7\xd4\xf1\x0e\x2b"
"\x48\x65\x82\x9f\xbc\xd9\xf6\x13"
"\x30\x4d\x6a\x87\xa4\xc1\xde\xfb"
"\x18\x35\x52\x6f\x8c\xa9\xc6\xe3"
"\x00\x1f\x3e\x5d\x7c\x9b\xba\xd9"
"\xf8\x17\x36\x55\x74\x93\xb2\xd1"
"\xf0\x0f\x2e\x4d\x6c\x8b\xaa\xc9"
"\xe8\x07\x26\x45\x64\x83\xa2\xc1"
"\xe0\xff\x1e\x3d\x5c\x7b\x9a\xb9"
"\xd8\xf7\x16\x35\x54\x73\x92\xb1"
"\xd0\xef\x0e\x2d\x4c\x6b\x8a\xa9"
"\xc8\xe7\x06\x25\x44\x63\x82\xa1"
"\xc0\xdf\xfe\x1d\x3c\x5b\x7a\x99"
"\xb8\xd7\xf6\x15\x34\x53\x72\x91"
"\xb0\xcf\xee\x0d\x2c\x4b\x6a\x89"
"\xa8\xc7\xe6\x05\x24\x43\x62\x81"
"\xa0\xbf\xde\xfd\x1c\x3b\x5a\x79"
"\x98\xb7\xd6\xf5\x14\x33\x52\x71"
"\x90\xaf\xce\xed\x0c\x2b\x4a\x69"
"\x88\xa7\xc6\xe5\x04\x23\x42\x61"
"\x80\x9f\xbe\xdd\xfc\x1b\x3a\x59"
"\x78\x97\xb6\xd5\xf4\x13\x32\x51"
"\x70\x8f\xae\xcd\xec\x0b\x2a\x49"
"\x68\x87\xa6\xc5\xe4\x03\x22\x41"
"\x60\x7f\x9e\xbd\xdc\xfb\x1a\x39"
"\x58\x77\x96\xb5\xd4\xf3\x12\x31"
"\x50\x6f\x8e\xad\xcc\xeb\x0a\x29"
"\x48\x67\x86\xa5\xc4\xe3\x02\x21"
"\x40\x5f\x7e\x9d\xbc\xdb\xfa\x19"
"\x38\x57\x76\x95\xb4\xd3\xf2\x11"
"\x30\x4f\x6e\x8d\xac\xcb\xea\x09"
"\x28\x47\x66\x85\xa4\xc3\xe2\x01"
"\x20\x3f\x5e\x7d\x9c\xbb\xda\xf9"
"\x18\x37\x56\x75\x94\xb3\xd2\xf1"
"\x10\x2f\x4e\x6d\x8c\xab\xca\xe9"
"\x08\x27\x46\x65\x84\xa3\xc2\xe1"
"\x00\x21\x42\x63",
.ctext =
"\xf0\x5c\x74\xad\x4e\xbc\x99\xe2"
"\xae\xff\x91\x3a\x44\xcf\x38\x32"
"\x1e\xad\xa7\xcd\xa1\x39\x95\xaa"
"\x10\xb1\xb3\x2e\x04\x31\x8f\x86"
"\xf2\x62\x74\x70\x0c\xa4\x46\x08"
"\xa8\xb7\x99\xa8\xe9\xd2\x73\x79"
"\x7e\x6e\xd4\x8f\x1e\xc7\x8e\x31"
"\x0b\xfa\x4b\xce\xfd\xf3\x57\x71"
"\xe9\x46\x03\xa5\x3d\x34\x00\xe2"
"\x18\xff\x75\x6d\x06\x2d\x00\xab"
"\xb9\x3e\x6c\x59\xc5\x84\x06\xb5"
"\x8b\xd0\x89\x9c\x4a\x79\x16\xc6"
"\x3d\x74\x54\xfa\x44\xcd\x23\x26"
"\x5c\xcf\x7e\x28\x92\x32\xbf\xdf"
"\xa7\x20\x3c\x74\x58\x2a\x9a\xde"
"\x61\x00\x1c\x4f\xff\x59\xc4\x22"
"\xac\x3c\xd0\xe8\x6c\xf9\x97\x1b"
"\x58\x9b\xad\x71\xe8\xa9\xb5\x0d"
"\xee\x2f\x04\x1f\x7f\xbc\x99\xee"
"\x84\xff\x42\x60\xdc\x3a\x18\xa5"
"\x81\xf9\xef\xdc\x7a\x0f\x65\x41"
"\x2f\xa3\xd3\xf9\xc2\xcb\xc0\x4d"
"\x8f\xd3\x76\x96\xad\x49\x6d\x38"
"\x3d\x39\x0b\x6c\x80\xb7\x54\x69"
"\xf0\x2c\x90\x02\x29\x0d\x1c\x12"
"\xad\x55\xc3\x8b\x68\xd9\xcc\xb3"
"\xb2\x64\x33\x90\x5e\xca\x4b\xe2"
"\xfb\x75\xdc\x63\xf7\x9f\x82\x74"
"\xf0\xc9\xaa\x7f\xe9\x2a\x9b\x33"
"\xbc\x88\x00\x7f\xca\xb2\x1f\x14"
"\xdb\xc5\x8e\x7b\x11\x3c\x3e\x08"
"\xf3\x83\xe8\xe0\x94\x86\x2e\x92"
"\x78\x6b\x01\xc9\xc7\x83\xba\x21"
"\x6a\x25\x15\x33\x4e\x45\x08\xec"
"\x35\xdb\xe0\x6e\x31\x51\x79\xa9"
"\x42\x44\x65\xc1\xa0\xf1\xf9\x2a"
"\x70\xd5\xb6\xc6\xc1\x8c\x39\xfc"
"\x25\xa6\x55\xd9\xdd\x2d\x4c\xec"
"\x49\xc6\xeb\x0e\xa8\x25\x2a\x16"
"\x1b\x66\x84\xda\xe2\x92\xe5\xc0"
"\xc8\x53\x07\xaf\x80\x84\xec\xfd"
"\xcd\xd1\x6e\xcd\x6f\x6a\xf5\x36"
"\xc5\x15\xe5\x25\x7d\x77\xd1\x1a"
"\x93\x36\xa9\xcf\x7c\xa4\x54\x4a"
"\x06\x51\x48\x4e\xf6\x59\x87\xd2"
"\x04\x02\xef\xd3\x44\xde\x76\x31"
"\xb3\x34\x17\x1b\x9d\x66\x11\x9f"
"\x1e\xcc\x17\xe9\xc7\x3c\x1b\xe7"
"\xcb\x50\x08\xfc\xdc\x2b\x24\xdb"
"\x65\x83\xd0\x3b\xe3\x30\xea\x94"
"\x6c\xe7\xe8\x35\x32\xc7\xdb\x64"
"\xb4\x01\xab\x36\x2c\x77\x13\xaf"
"\xf8\x2b\x88\x3f\x54\x39\xc4\x44"
"\xfe\xef\x6f\x68\x34\xbe\x0f\x05"
"\x16\x6d\xf6\x0a\x30\xe7\xe3\xed"
"\xc4\xde\x3c\x1b\x13\xd8\xdb\xfe"
"\x41\x62\xe5\x28\xd4\x8d\xa3\xc7"
"\x93\x97\xc6\x48\x45\x1d\x9f\x83"
"\xdf\x4b\x40\x3e\x42\x25\x87\x80"
"\x4c\x7d\xa8\xd4\x98\x23\x95\x75"
"\x41\x8c\xda\x41\x9b\xd4\xa7\x06"
"\xb5\xf1\x71\x09\x53\xbe\xca\xbf"
"\x32\x03\xed\xf0\x50\x1c\x56\x39"
"\x5b\xa4\x75\x18\xf7\x9b\x58\xef"
"\x53\xfc\x2a\x38\x23\x15\x75\xcd"
"\x45\xe5\x5a\x82\x55\xba\x21\xfa"
"\xd4\xbd\xc6\x94\x7c\xc5\x80\x12"
"\xf7\x4b\x32\xc4\x9a\x82\xd8\x28"
"\x8f\xd9\xc2\x0f\x60\x03\xbe\x5e"
"\x21\xd6\x5f\x58\xbf\x5c\xb1\x32"
"\x82\x8d\xa9\xe5\xf2\x66\x1a\xc0"
"\xa0\xbc\x58\x2f\x71\xf5\x2f\xed"
"\xd1\x26\xb9\xd8\x49\x5a\x07\x19"
"\x01\x7c\x59\xb0\xf8\xa4\xb7\xd3"
"\x7b\x1a\x8c\x38\xf4\x50\xa4\x59"
"\xb0\xcc\x41\x0b\x88\x7f\xe5\x31"
"\xb3\x42\xba\xa2\x7e\xd4\x32\x71"
"\x45\x87\x48\xa9\xc2\xf2\x89\xb3"
"\xe4\xa7\x7e\x52\x15\x61\xfa\xfe"
"\xc9\xdd\x81\xeb\x13\xab\xab\xc3"
"\x98\x59\xd8\x16\x3d\x14\x7a\x1c"
"\x3c\x41\x9a\x16\x16\x9b\xd2\xd2"
"\x69\x3a\x29\x23\xac\x86\x32\xa5"
"\x48\x9c\x9e\xf3\x47\x77\x81\x70"
"\x24\xe8\x85\xd2\xf5\xb5\xfa\xff"
"\x59\x6a\xd3\x50\x59\x43\x59\xde"
"\xd9\xf1\x55\xa5\x0c\xc3\x1a\x1a"
"\x18\x34\x0d\x1a\x63\x33\xed\x10"
"\xe0\x1d\x2a\x18\xd2\xc0\x54\xa8"
"\xca\xb5\x9a\xd3\xdd\xca\x45\x84"
"\x50\xe7\x0f\xfe\xa4\x99\x5a\xbe"
"\x43\x2d\x9a\xcb\x92\x3f\x5a\x1d"
"\x85\xd8\xc9\xdf\x68\xc9\x12\x80"
"\x56\x0c\xdc\x00\xdc\x3a\x7d\x9d"
"\xa3\xa2\xe8\x4d\xbf\xf9\x70\xa0"
"\xa4\x13\x4f\x6b\xaf\x0a\x89\x7f"
"\xda\xf0\xbf\x9b\xc8\x1d\xe5\xf8"
"\x2e\x8b\x07\xb5\x73\x1b\xcc\xa2"
"\xa6\xad\x30\xbc\x78\x3c\x5b\x10"
"\xfa\x5e\x62\x2d\x9e\x64\xb3\x33"
"\xce\xf9\x1f\x86\xe7\x8b\xa2\xb8"
"\xe8\x99\x57\x8c\x11\xed\x66\xd9"
"\x3c\x72\xb9\xc3\xe6\x4e\x17\x3a"
"\x6a\xcb\x42\x24\x06\xed\x3e\x4e"
"\xa3\xe8\x6a\x94\xda\x0d\x4e\xd5"
"\x14\x19\xcf\xb6\x26\xd8\x2e\xcc"
"\x64\x76\x38\x49\x4d\xfe\x30\x6d"
"\xe4\xc8\x8c\x7b\xc4\xe0\x35\xba"
"\x22\x6e\x76\xe1\x1a\xf2\x53\xc3"
"\x28\xa2\x82\x1f\x61\x69\xad\xc1"
"\x7b\x28\x4b\x1e\x6c\x85\x95\x9b"
"\x51\xb5\x17\x7f\x12\x69\x8c\x24"
"\xd5\xc7\x5a\x5a\x11\x54\xff\x5a"
"\xf7\x16\xc3\x91\xa6\xf0\xdc\x0a"
"\xb6\xa7\x4a\x0d\x7a\x58\xfe\xa5"
"\xf5\xcb\x8f\x7b\x0e\xea\x57\xe7"
"\xbd\x79\xd6\x1c\x88\x23\x6c\xf2"
"\x4d\x29\x77\x53\x35\x6a\x00\x8d"
"\xcd\xa3\x58\xbe\x77\x99\x18\xf8"
"\xe6\xe1\x8f\xe9\x37\x8f\xe3\xe2"
"\x5a\x8a\x93\x25\xaf\xf3\x78\x80"
"\xbe\xa6\x1b\xc6\xac\x8b\x1c\x91"
"\x58\xe1\x9f\x89\x35\x9d\x1d\x21"
"\x29\x9f\xf4\x99\x02\x27\x0f\xa8"
"\x4f\x79\x94\x2b\x33\x2c\xda\xa2"
"\x26\x39\x83\x94\xef\x27\xd8\x53"
"\x8f\x66\x0d\xe4\x41\x7d\x34\xcd"
"\x43\x7c\x95\x0a\x53\xef\x66\xda"
"\x7e\x9b\xf3\x93\xaf\xd0\x73\x71"
"\xba\x40\x9b\x74\xf8\xd7\xd7\x41"
"\x6d\xaf\x72\x9c\x8d\x21\x87\x3c"
"\xfd\x0a\x90\xa9\x47\x96\x9e\xd3"
"\x88\xee\x73\xcf\x66\x2f\x52\x56"
"\x6d\xa9\x80\x4c\xe2\x6f\x62\x88"
"\x3f\x0e\x54\x17\x48\x80\x5d\xd3"
"\xc3\xda\x25\x3d\xa1\xc8\xcb\x9f"
"\x9b\x70\xb3\xa1\xeb\x04\x52\xa1"
"\xf2\x22\x0f\xfc\xc8\x18\xfa\xf9"
"\x85\x9c\xf1\xac\xeb\x0c\x02\x46"
"\x75\xd2\xf5\x2c\xe3\xd2\x59\x94"
"\x12\xf3\x3c\xfc\xd7\x92\xfa\x36"
"\xba\x61\x34\x38\x7c\xda\x48\x3e"
"\x08\xc9\x39\x23\x5e\x02\x2c\x1a"
"\x18\x7e\xb4\xd9\xfd\x9e\x40\x02"
"\xb1\x33\x37\x32\xe7\xde\xd6\xd0"
"\x7c\x58\x65\x4b\xf8\x34\x27\x9c"
"\x44\xb4\xbd\xe9\xe9\x4c\x78\x7d"
"\x4b\x9f\xce\xb1\xcd\x47\xa5\x37"
"\xe5\x6d\xbd\xb9\x43\x94\x0a\xd4"
"\xd6\xf9\x04\x5f\xb5\x66\x6c\x1a"
"\x35\x12\xe3\x36\x28\x27\x36\x58"
"\x01\x2b\x79\xe4\xba\x6d\x10\x7d"
"\x65\xdf\x84\x95\xf4\xd5\xb6\x8f"
"\x2b\x9f\x96\x00\x86\x60\xf0\x21"
"\x76\xa8\x6a\x8c\x28\x1c\xb3\x6b"
"\x97\xd7\xb6\x53\x2a\xcc\xab\x40"
"\x9d\x62\x79\x58\x52\xe6\x65\xb7"
"\xab\x55\x67\x9c\x89\x7c\x03\xb0"
"\x73\x59\xc5\x81\xf5\x18\x17\x5c"
"\x89\xf3\x78\x35\x44\x62\x78\x72"
"\xd0\x96\xeb\x31\xe7\x87\x77\x14"
"\x99\x51\xf2\x59\x26\x9e\xb5\xa6"
"\x45\xfe\x6e\xbd\x07\x4c\x94\x5a"
"\xa5\x7d\xfc\xf1\x2b\x77\xe2\xfe"
"\x17\xd4\x84\xa0\xac\xb5\xc7\xda"
"\xa9\x1a\xb6\xf3\x74\x11\xb4\x9d"
"\xfb\x79\x2e\x04\x2d\x50\x28\x83"
"\xbf\xc6\x52\xd3\x34\xd6\xe8\x7a"
"\xb6\xea\xe7\xa8\x6c\x15\x1e\x2c"
"\x57\xbc\x48\x4e\x5f\x5c\xb6\x92"
"\xd2\x49\x77\x81\x6d\x90\x70\xae"
"\x98\xa1\x03\x0d\x6b\xb9\x77\x14"
"\xf1\x4e\x23\xd3\xf8\x68\xbd\xc2"
"\xfe\x04\xb7\x5c\xc5\x17\x60\x8f"
"\x65\x54\xa4\x7a\x42\xdc\x18\x0d"
"\xb5\xcf\x0f\xd3\xc7\x91\x66\x1b"
"\x45\x42\x27\x75\x50\xe5\xee\xb8"
"\x7f\x33\x2c\xba\x4a\x92\x4d\x2c"
"\x3c\xe3\x0d\x80\x01\xba\x0d\x29"
"\xd8\x3c\xe9\x13\x16\x57\xe6\xea"
"\x94\x52\xe7\x00\x4d\x30\xb0\x0f"
"\x35\xb8\xb8\xa7\xb1\xb5\x3b\x44"
"\xe1\x2f\xfd\x88\xed\x43\xe7\x52"
"\x10\x93\xb3\x8a\x30\x6b\x0a\xf7"
"\x23\xc6\x50\x9d\x4a\xb0\xde\xc3"
"\xdc\x9b\x2f\x01\x56\x36\x09\xc5"
"\x2f\x6b\xfe\xf1\xd8\x27\x45\x03"
"\x30\x5e\x5c\x5b\xb4\x62\x0e\x1a"
"\xa9\x21\x2b\x92\x94\x87\x62\x57"
"\x4c\x10\x74\x1a\xf1\x0a\xc5\x84"
"\x3b\x9e\x72\x02\xd7\xcc\x09\x56"
"\xbd\x54\xc1\xf0\xc3\xe3\xb3\xf8"
"\xd2\x0d\x61\xcb\xef\xce\x0d\x05"
"\xb0\x98\xd9\x8e\x4f\xf9\xbc\x93"
"\xa6\xea\xc8\xcf\x10\x53\x4b\xf1"
"\xec\xfc\x89\xf9\x64\xb0\x22\xbf"
"\x9e\x55\x46\x9f\x7c\x50\x8e\x84"
"\x54\x20\x98\xd7\x6c\x40\x1e\xdb"
"\x69\x34\x78\x61\x24\x21\x9c\x8a"
"\xb3\x62\x31\x8b\x6e\xf5\x2a\x35"
"\x86\x13\xb1\x6c\x64\x2e\x41\xa5"
"\x05\xf2\x42\xba\xd2\x3a\x0d\x8e"
"\x8a\x59\x94\x3c\xcf\x36\x27\x82"
"\xc2\x45\xee\x58\xcd\x88\xb4\xec"
"\xde\xb2\x96\x0a\xaf\x38\x6f\x88"
"\xd7\xd8\xe1\xdf\xb9\x96\xa9\x0a"
"\xb1\x95\x28\x86\x20\xe9\x17\x49"
"\xa2\x29\x38\xaa\xa5\xe9\x6e\xf1"
"\x19\x27\xc0\xd5\x2a\x22\xc3\x0b"
"\xdb\x7c\x73\x10\xb9\xba\x89\x76"
"\x54\xae\x7d\x71\xb3\x93\xf6\x32"
"\xe6\x47\x43\x55\xac\xa0\x0d\xc2"
"\x93\x27\x4a\x8e\x0e\x74\x15\xc7"
"\x0b\x85\xd9\x0c\xa9\x30\x7a\x3e"
"\xea\x8f\x85\x6d\x3a\x12\x4f\x72"
"\x69\x58\x7a\x80\xbb\xb5\x97\xf3"
"\xcf\x70\xd2\x5d\xdd\x4d\x21\x79"
"\x54\x4d\xe4\x05\xe8\xbd\xc2\x62"
"\xb1\x3b\x77\x1c\xd6\x5c\xf3\xa0"
"\x79\x00\xa8\x6c\x29\xd9\x18\x24"
"\x36\xa2\x46\xc0\x96\x65\x7f\xbd"
"\x2a\xed\x36\x16\x0c\xaa\x9f\xf4"
"\xc5\xb4\xe2\x12\xed\x69\xed\x4f"
"\x26\x2c\x39\x52\x89\x98\xe7\x2c"
"\x99\xa4\x9e\xa3\x9b\x99\x46\x7a"
"\x3a\xdc\xa8\x59\xa3\xdb\xc3\x3b"
"\x95\x0d\x3b\x09\x6e\xee\x83\x5d"
"\x32\x4d\xed\xab\xfa\x98\x14\x4e"
"\xc3\x15\x45\x53\x61\xc4\x93\xbd"
"\x90\xf4\x99\x95\x4c\xe6\x76\x92"
"\x29\x90\x46\x30\x92\x69\x7d\x13"
"\xf2\xa5\xcd\x69\x49\x44\xb2\x0f"
"\x63\x40\x36\x5f\x09\xe2\x78\xf8"
"\x91\xe3\xe2\xfa\x10\xf7\xc8\x24"
"\xa8\x89\x32\x5c\x37\x25\x1d\xb2"
"\xea\x17\x8a\x0a\xa9\x64\xc3\x7c"
"\x3c\x7c\xbd\xc6\x79\x34\xe7\xe2"
"\x85\x8e\xbf\xf8\xde\x92\xa0\xae"
"\x20\xc4\xf6\xbb\x1f\x38\x19\x0e"
"\xe8\x79\x9c\xa1\x23\xe9\x54\x7e"
"\x37\x2f\xe2\x94\x32\xaf\xa0\x23"
"\x49\xe4\xc0\xb3\xac\x00\x8f\x36"
"\x05\xc4\xa6\x96\xec\x05\x98\x4f"
"\x96\x67\x57\x1f\x20\x86\x1b\x2d"
"\x69\xe4\x29\x93\x66\x5f\xaf\x6b"
"\x88\x26\x2c\x67\x02\x4b\x52\xd0"
"\x83\x7a\x43\x1f\xc0\x71\x15\x25"
"\x77\x65\x08\x60\x11\x76\x4c\x8d"
"\xed\xa9\x27\xc6\xb1\x2a\x2c\x6a"
"\x4a\x97\xf5\xc6\xb7\x70\x42\xd3"
"\x03\xd1\x24\x95\xec\x6d\xab\x38"
"\x72\xce\xe2\x8b\x33\xd7\x51\x09"
"\xdc\x45\xe0\x09\x96\x32\xf3\xc4"
"\x84\xdc\x73\x73\x2d\x1b\x11\x98"
"\xc5\x0e\x69\x28\x94\xc7\xb5\x4d"
"\xc8\x8a\xd0\xaa\x13\x2e\x18\x74"
"\xdd\xd1\x1e\xf3\x90\xe8\xfc\x9a"
"\x72\x4a\x0e\xd1\xe4\xfb\x0d\x96"
"\xd1\x0c\x79\x85\x1b\x1c\xfe\xe1"
"\x62\x8f\x7a\x73\x32\xab\xc8\x18"
"\x69\xe3\x34\x30\xdf\x13\xa6\xe5"
"\xe8\x0e\x67\x7f\x81\x11\xb4\x60"
"\xc7\xbd\x79\x65\x50\xdc\xc4\x5b"
"\xde\x39\xa4\x01\x72\x63\xf3\xd1"
"\x64\x4e\xdf\xfc\x27\x92\x37\x0d"
"\x57\xcd\x11\x4f\x11\x04\x8e\x1d"
"\x16\xf7\xcd\x92\x9a\x99\x30\x14"
"\xf1\x7c\x67\x1b\x1f\x41\x0b\xe8"
"\x32\xe8\xb8\xc1\x4f\x54\x86\x4f"
"\xe5\x79\x81\x73\xcd\x43\x59\x68"
"\x73\x02\x3b\x78\x21\x72\x43\x00"
"\x49\x17\xf7\x00\xaf\x68\x24\x53"
"\x05\x0a\xc3\x33\xe0\x33\x3f\x69"
"\xd2\x84\x2f\x0b\xed\xde\x04\xf4"
"\x11\x94\x13\x69\x51\x09\x28\xde"
"\x57\x5c\xef\xdc\x9a\x49\x1c\x17"
"\x97\xf3\x96\xc1\x7f\x5d\x2e\x7d"
"\x55\xb8\xb3\x02\x09\xb3\x1f\xe7"
"\xc9\x8d\xa3\x36\x34\x8a\x77\x13"
"\x30\x63\x4c\xa5\xcd\xc3\xe0\x7e"
"\x05\xa1\x7b\x0c\xcb\x74\x47\x31"
"\x62\x03\x43\xf1\x87\xb4\xb0\x85"
"\x87\x8e\x4b\x25\xc7\xcf\xae\x4b"
"\x36\x46\x3e\x62\xbc\x6f\xeb\x5f"
"\x73\xac\xe6\x07\xee\xc1\xa1\xd6"
"\xc4\xab\xc9\xd6\x89\x45\xe1\xf1"
"\x04\x4e\x1a\x6f\xbb\x4f\x3a\xa3"
"\xa0\xcb\xa3\x0a\xd8\x71\x35\x55"
"\xe4\xbc\x2e\x04\x06\xe6\xff\x5b"
"\x1c\xc0\x11\x7c\xc5\x17\xf3\x38"
"\xcf\xe9\xba\x0f\x0e\xef\x02\xc2"
"\x8d\xc6\xbc\x4b\x67\x20\x95\xd7"
"\x2c\x45\x5b\x86\x44\x8c\x6f\x2e"
"\x7e\x9f\x1c\x77\xba\x6b\x0e\xa3"
"\x69\xdc\xab\x24\x57\x60\x47\xc1"
"\xd1\xa5\x9d\x23\xe6\xb1\x37\xfe"
"\x93\xd2\x4c\x46\xf9\x0c\xc6\xfb"
"\xd6\x9d\x99\x69\xab\x7a\x07\x0c"
"\x65\xe7\xc4\x08\x96\xe2\xa5\x01"
"\x3f\x46\x07\x05\x7e\xe8\x9a\x90"
"\x50\xdc\xe9\x7a\xea\xa1\x39\x6e"
"\x66\xe4\x6f\xa5\x5f\xb2\xd9\x5b"
"\xf5\xdb\x2a\x32\xf0\x11\x6f\x7c"
"\x26\x10\x8f\x3d\x80\xe9\x58\xf7"
"\xe0\xa8\x57\xf8\xdb\x0e\xce\x99"
"\x63\x19\x3d\xd5\xec\x1b\x77\x69"
"\x98\xf6\xe4\x5f\x67\x17\x4b\x09"
"\x85\x62\x82\x70\x18\xe2\x9a\x78"
"\xe2\x62\xbd\xb4\xf1\x42\xc6\xfb"
"\x08\xd0\xbd\xeb\x4e\x09\xf2\xc8"
"\x1e\xdc\x3d\x32\x21\x56\x9c\x4f"
"\x35\xf3\x61\x06\x72\x84\xc4\x32"
"\xf2\xf1\xfa\x0b\x2f\xc3\xdb\x02"
"\x04\xc2\xde\x57\x64\x60\x8d\xcf"
"\xcb\x86\x5d\x97\x3e\xb1\x9c\x01"
"\xd6\x28\x8f\x99\xbc\x46\xeb\x05"
"\xaf\x7e\xb8\x21\x2a\x56\x85\x1c"
"\xb3\x71\xa0\xde\xca\x96\xf1\x78"
"\x49\xa2\x99\x81\x80\x5c\x01\xf5"
"\xa0\xa2\x56\x63\xe2\x70\x07\xa5"
"\x95\xd6\x85\xeb\x36\x9e\xa9\x51"
"\x66\x56\x5f\x1d\x02\x19\xe2\xf6"
"\x4f\x73\x38\x09\x75\x64\x48\xe0"
"\xf1\x7e\x0e\xe8\x9d\xf9\xed\x94"
"\xfe\x16\x26\x62\x49\x74\xf4\xb0"
"\xd4\xa9\x6c\xb0\xfd\x53\xe9\x81"
"\xe0\x7a\xbf\xcf\xb5\xc4\x01\x81"
"\x79\x99\x77\x01\x3b\xe9\xa2\xb6"
"\xe6\x6a\x8a\x9e\x56\x1c\x8d\x1e"
"\x8f\x06\x55\x2c\x6c\xdc\x92\x87"
"\x64\x3b\x4b\x19\xa1\x13\x64\x1d"
"\x4a\xe9\xc0\x00\xb8\x95\xef\x6b"
"\x1a\x86\x6d\x37\x52\x02\xc2\xe0"
"\xc8\xbb\x42\x0c\x02\x21\x4a\xc9"
"\xef\xa0\x54\xe4\x5e\x16\x53\x81"
"\x70\x62\x10\xaf\xde\xb8\xb5\xd3"
"\xe8\x5e\x6c\xc3\x8a\x3e\x18\x07"
"\xf2\x2f\x7d\xa7\xe1\x3d\x4e\xb4"
"\x26\xa7\xa3\x93\x86\xb2\x04\x1e"
"\x53\x5d\x86\xd6\xde\x65\xca\xe3"
"\x4e\xc1\xcf\xef\xc8\x70\x1b\x83"
"\x13\xdd\x18\x8b\x0d\x76\xd2\xf6"
"\x37\x7a\x93\x7a\x50\x11\x9f\x96"
"\x86\x25\xfd\xac\xdc\xbe\x18\x93"
"\x19\x6b\xec\x58\x4f\xb9\x75\xa7"
"\xdd\x3f\x2f\xec\xc8\x5a\x84\xab"
"\xd5\xe4\x8a\x07\xf6\x4d\x23\xd6"
"\x03\xfb\x03\x6a\xea\x66\xbf\xd4"
"\xb1\x34\xfb\x78\xe9\x55\xdc\x7c"
"\x3d\x9c\xe5\x9a\xac\xc3\x7a\x80"
"\x24\x6d\xa0\xef\x25\x7c\xb7\xea"
"\xce\x4d\x5f\x18\x60\xce\x87\x22"
"\x66\x2f\xd5\xdd\xdd\x02\x21\x75"
"\x82\xa0\x1f\x58\xc6\xd3\x62\xf7"
"\x32\xd8\xaf\x1e\x07\x77\x51\x96"
"\xd5\x6b\x1e\x7e\x80\x02\xe8\x67"
"\xea\x17\x0b\x10\xd2\x3f\x28\x25"
"\x4f\x05\x77\x02\x14\x69\xf0\x2c"
"\xbe\x0c\xf1\x74\x30\xd1\xb9\x9b"
"\xfc\x8c\xbb\x04\x16\xd9\xba\xc3"
"\xbc\x91\x8a\xc4\x30\xa4\xb0\x12"
"\x4c\x21\x87\xcb\xc9\x1d\x16\x96"
"\x07\x6f\x23\x54\xb9\x6f\x79\xe5"
"\x64\xc0\x64\xda\xb1\xae\xdd\x60"
"\x6c\x1a\x9d\xd3\x04\x8e\x45\xb0"
"\x92\x61\xd0\x48\x81\xed\x5e\x1d"
"\xa0\xc9\xa4\x33\xc7\x13\x51\x5d"
"\x7f\x83\x73\xb6\x70\x18\x65\x3e"
"\x2f\x0e\x7a\x12\x39\x98\xab\xd8"
"\x7e\x6f\xa3\xd1\xba\x56\xad\xbd"
"\xf0\x03\x01\x1c\x85\x35\x9f\xeb"
"\x19\x63\xa1\xaf\xfe\x2d\x35\x50"
"\x39\xa0\x65\x7c\x95\x7e\x6b\xfe"
"\xc1\xac\x07\x7c\x98\x4f\xbe\x57"
"\xa7\x22\xec\xe2\x7e\x29\x09\x53"
"\xe8\xbf\xb4\x7e\x3f\x8f\xfc\x14"
"\xce\x54\xf9\x18\x58\xb5\xff\x44"
"\x05\x9d\xce\x1b\xb6\x82\x23\xc8"
"\x2e\xbc\x69\xbb\x4a\x29\x0f\x65"
"\x94\xf0\x63\x06\x0e\xef\x8c\xbd"
"\xff\xfd\xb0\x21\x6e\x57\x05\x75"
"\xda\xd5\xc4\xeb\x8d\x32\xf7\x50"
"\xd3\x6f\x22\xed\x5f\x8e\xa2\x5b"
"\x80\x8c\xc8\x78\x40\x24\x4b\x89"
"\x30\xce\x7a\x97\x0e\xc4\xaf\xef"
"\x9b\xb4\xcd\x66\x74\x14\x04\x2b"
"\xf7\xce\x0b\x1c\x6e\xc2\x78\x8c"
"\xca\xc5\xd0\x1c\x95\x4a\x91\x2d"
"\xa7\x20\xeb\x86\x52\xb7\x67\xd8"
"\x0c\xd6\x04\x14\xde\x51\x74\x75"
"\xe7\x11\xb4\x87\xa3\x3d\x2d\xad"
"\x4f\xef\xa0\x0f\x70\x00\x6d\x13"
"\x19\x1d\x41\x50\xe9\xd8\xf0\x32"
"\x71\xbc\xd3\x11\xf2\xac\xbe\xaf"
"\x75\x46\x65\x4e\x07\x34\x37\xa3"
"\x89\xfe\x75\xd4\x70\x4c\xc6\x3f"
"\x69\x24\x0e\x38\x67\x43\x8c\xde"
"\x06\xb5\xb8\xe7\xc4\xf0\x41\x8f"
"\xf0\xbd\x2f\x0b\xb9\x18\xf8\xde"
"\x64\xb1\xdb\xee\x00\x50\x77\xe1"
"\xc7\xff\xa6\xfa\xdd\x70\xf4\xe3"
"\x93\xe9\x77\x35\x3d\x4b\x2f\x2b"
"\x6d\x55\xf0\xfc\x88\x54\x4e\x89"
"\xc1\x8a\x23\x31\x2d\x14\x2a\xb8"
"\x1b\x15\xdd\x9e\x6e\x7b\xda\x05"
"\x91\x7d\x62\x64\x96\x72\xde\xfc"
"\xc1\xec\xf0\x23\x51\x6f\xdb\x5b"
"\x1d\x08\x57\xce\x09\xb8\xf6\xcd"
"\x8d\x95\xf2\x20\xbf\x0f\x20\x57"
"\x98\x81\x84\x4f\x15\x5c\x76\xe7"
"\x3e\x0a\x3a\x6c\xc4\x8a\xbe\x78"
"\x74\x77\xc3\x09\x4b\x5d\x48\xe4"
"\xc8\xcb\x0b\xea\x17\x28\xcf\xcf"
"\x31\x32\x44\xa4\xe5\x0e\x1a\x98"
"\x94\xc4\xf0\xff\xae\x3e\x44\xe8"
"\xa5\xb3\xb5\x37\x2f\xe8\xaf\x6f"
"\x28\xc1\x37\x5f\x31\xd2\xb9\x33"
"\xb1\xb2\x52\x94\x75\x2c\x29\x59"
"\x06\xc2\x25\xe8\x71\x65\x4e\xed"
"\xc0\x9c\xb1\xbb\x25\xdc\x6c\xe7"
"\x4b\xa5\x7a\x54\x7a\x60\xff\x7a"
"\xe0\x50\x40\x96\x35\x63\xe4\x0b"
"\x76\xbd\xa4\x65\x00\x1b\x57\x88"
"\xae\xed\x39\x88\x42\x11\x3c\xed"
"\x85\x67\x7d\xb9\x68\x82\xe9\x43"
"\x3c\x47\x53\xfa\xe8\xf8\x9f\x1f"
"\x9f\xef\x0f\xf7\x30\xd9\x30\x0e"
"\xb9\x9f\x69\x18\x2f\x7e\xf8\xf8"
"\xf8\x8c\x0f\xd4\x02\x4d\xea\xcd"
"\x0a\x9c\x6f\x71\x6d\x5a\x4c\x60"
"\xce\x20\x56\x32\xc6\xc5\x99\x1f"
"\x09\xe6\x4e\x18\x1a\x15\x13\xa8"
"\x7d\xb1\x6b\xc0\xb2\x6d\xf8\x26"
"\x66\xf8\x3d\x18\x74\x70\x66\x7a"
"\x34\x17\xde\xba\x47\xf1\x06\x18"
"\xcb\xaf\xeb\x4a\x1e\x8f\xa7\x77"
"\xe0\x3b\x78\x62\x66\xc9\x10\xea"
"\x1f\xb7\x29\x0a\x45\xa1\x1d\x1e"
"\x1d\xe2\x65\x61\x50\x9c\xd7\x05"
"\xf2\x0b\x5b\x12\x61\x02\xc8\xe5"
"\x63\x4f\x20\x0c\x07\x17\x33\x5e"
"\x03\x9a\x53\x0f\x2e\x55\xfe\x50"
"\x43\x7d\xd0\xb6\x7e\x5a\xda\xae"
"\x58\xef\x15\xa9\x83\xd9\x46\xb1"
"\x42\xaa\xf5\x02\x6c\xce\x92\x06"
"\x1b\xdb\x66\x45\x91\x79\xc2\x2d"
"\xe6\x53\xd3\x14\xfd\xbb\x44\x63"
"\xc6\xd7\x3d\x7a\x0c\x75\x78\x9d"
"\x5c\xa6\x39\xb3\xe5\x63\xca\x8b"
"\xfe\xd3\xef\x60\x83\xf6\x8e\x70"
"\xb6\x67\xc7\x77\xed\x23\xef\x4c"
"\xf0\xed\x2d\x07\x59\x6f\xc1\x01"
"\x34\x37\x08\xab\xd9\x1f\x09\xb1"
"\xce\x5b\x17\xff\x74\xf8\x9c\xd5"
"\x2c\x56\x39\x79\x0f\x69\x44\x75"
"\x58\x27\x01\xc4\xbf\xa7\xa1\x1d"
"\x90\x17\x77\x86\x5a\x3f\xd9\xd1"
"\x0e\xa0\x10\xf8\xec\x1e\xa5\x7f"
"\x5e\x36\xd1\xe3\x04\x2c\x70\xf7"
"\x8e\xc0\x98\x2f\x6c\x94\x2b\x41"
"\xb7\x60\x00\xb7\x2e\xb8\x02\x8d"
"\xb8\xb0\xd3\x86\xba\x1d\xd7\x90"
"\xd6\xb6\xe1\xfc\xd7\xd8\x28\x06"
"\x63\x9b\xce\x61\x24\x79\xc0\x70"
"\x52\xd0\xb6\xd4\x28\x95\x24\x87"
"\x03\x1f\xb7\x9a\xda\xa3\xfb\x52"
"\x5b\x68\xe7\x4c\x8c\x24\xe1\x42"
"\xf7\xd5\xfd\xad\x06\x32\x9f\xba"
"\xc1\xfc\xdd\xc6\xfc\xfc\xb3\x38"
"\x74\x56\x58\x40\x02\x37\x52\x2c"
"\x55\xcc\xb3\x9e\x7a\xe9\xd4\x38"
"\x41\x5e\x0c\x35\xe2\x11\xd1\x13"
"\xf8\xb7\x8d\x72\x6b\x22\x2a\xb0"
"\xdb\x08\xba\x35\xb9\x3f\xc8\xd3"
"\x24\x90\xec\x58\xd2\x09\xc7\x2d"
"\xed\x38\x80\x36\x72\x43\x27\x49"
"\x4a\x80\x8a\xa2\xe8\xd3\xda\x30"
"\x7d\xb6\x82\x37\x86\x92\x86\x3e"
"\x08\xb2\x28\x5a\x55\x44\x24\x7d"
"\x40\x48\x8a\xb6\x89\x58\x08\xa0"
"\xd6\x6d\x3a\x17\xbf\xf6\x54\xa2"
"\xf5\xd3\x8c\x0f\x78\x12\x57\x8b"
"\xd5\xc2\xfd\x58\x5b\x7f\x38\xe3"
"\xcc\xb7\x7c\x48\xb3\x20\xe8\x81"
"\x14\x32\x45\x05\xe0\xdb\x9f\x75"
"\x85\xb4\x6a\xfc\x95\xe3\x54\x22"
"\x12\xee\x30\xfe\xd8\x30\xef\x34"
"\x50\xab\x46\x30\x98\x2f\xb7\xc0"
"\x15\xa2\x83\xb6\xf2\x06\x21\xa2"
"\xc3\x26\x37\x14\xd1\x4d\xb5\x10"
"\x52\x76\x4d\x6a\xee\xb5\x2b\x15"
"\xb7\xf9\x51\xe8\x2a\xaf\xc7\xfa"
"\x77\xaf\xb0\x05\x4d\xd1\x68\x8e"
"\x74\x05\x9f\x9d\x93\xa5\x3e\x7f"
"\x4e\x5f\x9d\xcb\x09\xc7\x83\xe3"
"\x02\x9d\x27\x1f\xef\x85\x05\x8d"
"\xec\x55\x88\x0f\x0d\x7c\x4c\xe8"
"\xa1\x75\xa0\xd8\x06\x47\x14\xef"
"\xaa\x61\xcf\x26\x15\xad\xd8\xa3"
"\xaa\x75\xf2\x78\x4a\x5a\x61\xdf"
"\x8b\xc7\x04\xbc\xb2\x32\xd2\x7e"
"\x42\xee\xb4\x2f\x51\xff\x7b\x2e"
"\xd3\x02\xe8\xdc\x5d\x0d\x50\xdc"
"\xae\xb7\x46\xf9\xa8\xe6\xd0\x16"
"\xcc\xe6\x2c\x81\xc7\xad\xe9\xf0"
"\x05\x72\x6d\x3d\x0a\x7a\xa9\x02"
"\xac\x82\x93\x6e\xb6\x1c\x28\xfc"
"\x44\x12\xfb\x73\x77\xd4\x13\x39"
"\x29\x88\x8a\xf3\x5c\xa6\x36\xa0"
"\x2a\xed\x7e\xb1\x1d\xd6\x4c\x6b"
"\x41\x01\x18\x5d\x5d\x07\x97\xa6"
"\x4b\xef\x31\x18\xea\xac\xb1\x84"
"\x21\xed\xda\x86",
.len = 4100,
.np = 2,
.tap = { 4064, 36 },
crypto: testmgr - Add ChaCha20-Poly1305 test vectors from RFC7539 Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:44:01 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec aes_ofb_tv_template[] = {
/* From NIST Special Publication 800-38A, Appendix F.5 */
{
.key = "\x2b\x7e\x15\x16\x28\xae\xd2\xa6"
"\xab\xf7\x15\x88\x09\xcf\x4f\x3c",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.ptext = "\x6b\xc1\xbe\xe2\x2e\x40\x9f\x96"
"\xe9\x3d\x7e\x11\x73\x93\x17\x2a"
"\xae\x2d\x8a\x57\x1e\x03\xac\x9c"
"\x9e\xb7\x6f\xac\x45\xaf\x8e\x51"
"\x30\xc8\x1c\x46\xa3\x5c\xe4\x11"
"\xe5\xfb\xc1\x19\x1a\x0a\x52\xef"
"\xf6\x9f\x24\x45\xdf\x4f\x9b\x17"
"\xad\x2b\x41\x7b\xe6\x6c\x37\x10",
.ctext = "\x3b\x3f\xd9\x2e\xb7\x2d\xad\x20"
"\x33\x34\x49\xf8\xe8\x3c\xfb\x4a"
"\x77\x89\x50\x8d\x16\x91\x8f\x03\xf5"
"\x3c\x52\xda\xc5\x4e\xd8\x25"
"\x97\x40\x05\x1e\x9c\x5f\xec\xf6\x43"
"\x44\xf7\xa8\x22\x60\xed\xcc"
"\x30\x4c\x65\x28\xf6\x59\xc7\x78"
"\x66\xa5\x10\xd9\xc1\xd6\xae\x5e",
.len = 64,
}
};
static const struct aead_testvec aes_gcm_enc_tv_template[] = {
{ /* From McGrew & Viega - http://citeseer.ist.psu.edu/656989.html */
.key = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x58\xe2\xfc\xce\xfa\x7e\x30\x61"
"\x36\x7f\x1d\x57\xa4\xe7\x45\x5a",
.rlen = 16,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x03\x88\xda\xce\x60\xb6\xa3\x92"
"\xf3\x28\xc2\xb9\x71\xb2\xfe\x78"
"\xab\x6e\x47\xd4\x2c\xec\x13\xbd"
"\xf5\x3a\x67\xb2\x12\x57\xbd\xdf",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39\x1a\xaf\xd2\x55",
.ilen = 64,
.result = "\x42\x83\x1e\xc2\x21\x77\x74\x24"
"\x4b\x72\x21\xb7\x84\xd0\xd4\x9c"
"\xe3\xaa\x21\x2f\x2c\x02\xa4\xe0"
"\x35\xc1\x7e\x23\x29\xac\xa1\x2e"
"\x21\xd5\x14\xb2\x54\x66\x93\x1c"
"\x7d\x8f\x6a\x5a\xac\x84\xaa\x05"
"\x1b\xa3\x0b\x39\x6a\x0a\xac\x97"
"\x3d\x58\xe0\x91\x47\x3f\x59\x85"
"\x4d\x5c\x2a\xf3\x27\xcd\x64\xa6"
"\x2c\xf3\x5a\xbd\x2b\xa6\xfa\xb4",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39",
.ilen = 60,
.assoc = "\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xab\xad\xda\xd2",
.alen = 20,
.result = "\x42\x83\x1e\xc2\x21\x77\x74\x24"
"\x4b\x72\x21\xb7\x84\xd0\xd4\x9c"
"\xe3\xaa\x21\x2f\x2c\x02\xa4\xe0"
"\x35\xc1\x7e\x23\x29\xac\xa1\x2e"
"\x21\xd5\x14\xb2\x54\x66\x93\x1c"
"\x7d\x8f\x6a\x5a\xac\x84\xaa\x05"
"\x1b\xa3\x0b\x39\x6a\x0a\xac\x97"
"\x3d\x58\xe0\x91"
"\x5b\xc9\x4f\xbc\x32\x21\xa5\xdb"
"\x94\xfa\xe9\x5a\xe7\x12\x1a\x47",
.rlen = 76,
}, {
.key = zeroed_string,
.klen = 24,
.result = "\xcd\x33\xb2\x8a\xc7\x73\xf7\x4b"
"\xa0\x0e\xd1\xf3\x12\x57\x24\x35",
.rlen = 16,
}, {
.key = zeroed_string,
.klen = 24,
.input = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x98\xe7\x24\x7c\x07\xf0\xfe\x41"
"\x1c\x26\x7e\x43\x84\xb0\xf6\x00"
"\x2f\xf5\x8d\x80\x03\x39\x27\xab"
"\x8e\xf4\xd4\x58\x75\x14\xf0\xfb",
.rlen = 32,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c",
.klen = 24,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39\x1a\xaf\xd2\x55",
.ilen = 64,
.result = "\x39\x80\xca\x0b\x3c\x00\xe8\x41"
"\xeb\x06\xfa\xc4\x87\x2a\x27\x57"
"\x85\x9e\x1c\xea\xa6\xef\xd9\x84"
"\x62\x85\x93\xb4\x0c\xa1\xe1\x9c"
"\x7d\x77\x3d\x00\xc1\x44\xc5\x25"
"\xac\x61\x9d\x18\xc8\x4a\x3f\x47"
"\x18\xe2\x44\x8b\x2f\xe3\x24\xd9"
"\xcc\xda\x27\x10\xac\xad\xe2\x56"
"\x99\x24\xa7\xc8\x58\x73\x36\xbf"
"\xb1\x18\x02\x4d\xb8\x67\x4a\x14",
.rlen = 80,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c",
.klen = 24,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39",
.ilen = 60,
.assoc = "\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xab\xad\xda\xd2",
.alen = 20,
.result = "\x39\x80\xca\x0b\x3c\x00\xe8\x41"
"\xeb\x06\xfa\xc4\x87\x2a\x27\x57"
"\x85\x9e\x1c\xea\xa6\xef\xd9\x84"
"\x62\x85\x93\xb4\x0c\xa1\xe1\x9c"
"\x7d\x77\x3d\x00\xc1\x44\xc5\x25"
"\xac\x61\x9d\x18\xc8\x4a\x3f\x47"
"\x18\xe2\x44\x8b\x2f\xe3\x24\xd9"
"\xcc\xda\x27\x10"
"\x25\x19\x49\x8e\x80\xf1\x47\x8f"
"\x37\xba\x55\xbd\x6d\x27\x61\x8c",
.rlen = 76,
.np = 2,
.tap = { 32, 28 },
.anp = 2,
.atap = { 8, 12 }
}, {
.key = zeroed_string,
.klen = 32,
.result = "\x53\x0f\x8a\xfb\xc7\x45\x36\xb9"
"\xa9\x63\xb4\xf1\xc4\xcb\x73\x8b",
.rlen = 16,
}
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aes_gcm_dec_tv_template[] = {
{ /* From McGrew & Viega - http://citeseer.ist.psu.edu/656989.html */
.key = zeroed_string,
.klen = 32,
.input = "\xce\xa7\x40\x3d\x4d\x60\x6b\x6e"
"\x07\x4e\xc5\xd3\xba\xf3\x9d\x18"
"\xd0\xd1\xc8\xa7\x99\x99\x6b\xf0"
"\x26\x5b\x98\xb5\xd4\x8a\xb9\x19",
.ilen = 32,
.result = zeroed_string,
.rlen = 16,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
.klen = 32,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x52\x2d\xc1\xf0\x99\x56\x7d\x07"
"\xf4\x7f\x37\xa3\x2a\x84\x42\x7d"
"\x64\x3a\x8c\xdc\xbf\xe5\xc0\xc9"
"\x75\x98\xa2\xbd\x25\x55\xd1\xaa"
"\x8c\xb0\x8e\x48\x59\x0d\xbb\x3d"
"\xa7\xb0\x8b\x10\x56\x82\x88\x38"
"\xc5\xf6\x1e\x63\x93\xba\x7a\x0a"
"\xbc\xc9\xf6\x62\x89\x80\x15\xad"
"\xb0\x94\xda\xc5\xd9\x34\x71\xbd"
"\xec\x1a\x50\x22\x70\xe3\xcc\x6c",
.ilen = 80,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39\x1a\xaf\xd2\x55",
.rlen = 64,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
.klen = 32,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x52\x2d\xc1\xf0\x99\x56\x7d\x07"
"\xf4\x7f\x37\xa3\x2a\x84\x42\x7d"
"\x64\x3a\x8c\xdc\xbf\xe5\xc0\xc9"
"\x75\x98\xa2\xbd\x25\x55\xd1\xaa"
"\x8c\xb0\x8e\x48\x59\x0d\xbb\x3d"
"\xa7\xb0\x8b\x10\x56\x82\x88\x38"
"\xc5\xf6\x1e\x63\x93\xba\x7a\x0a"
"\xbc\xc9\xf6\x62"
"\x76\xfc\x6e\xce\x0f\x4e\x17\x68"
"\xcd\xdf\x88\x53\xbb\x2d\x55\x1b",
.ilen = 76,
.assoc = "\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xab\xad\xda\xd2",
.alen = 20,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39",
.rlen = 60,
.np = 2,
.tap = { 48, 28 },
.anp = 3,
.atap = { 8, 8, 4 }
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x42\x83\x1e\xc2\x21\x77\x74\x24"
"\x4b\x72\x21\xb7\x84\xd0\xd4\x9c"
"\xe3\xaa\x21\x2f\x2c\x02\xa4\xe0"
"\x35\xc1\x7e\x23\x29\xac\xa1\x2e"
"\x21\xd5\x14\xb2\x54\x66\x93\x1c"
"\x7d\x8f\x6a\x5a\xac\x84\xaa\x05"
"\x1b\xa3\x0b\x39\x6a\x0a\xac\x97"
"\x3d\x58\xe0\x91\x47\x3f\x59\x85"
"\x4d\x5c\x2a\xf3\x27\xcd\x64\xa6"
"\x2c\xf3\x5a\xbd\x2b\xa6\xfa\xb4",
.ilen = 80,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39\x1a\xaf\xd2\x55",
.rlen = 64,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x42\x83\x1e\xc2\x21\x77\x74\x24"
"\x4b\x72\x21\xb7\x84\xd0\xd4\x9c"
"\xe3\xaa\x21\x2f\x2c\x02\xa4\xe0"
"\x35\xc1\x7e\x23\x29\xac\xa1\x2e"
"\x21\xd5\x14\xb2\x54\x66\x93\x1c"
"\x7d\x8f\x6a\x5a\xac\x84\xaa\x05"
"\x1b\xa3\x0b\x39\x6a\x0a\xac\x97"
"\x3d\x58\xe0\x91"
"\x5b\xc9\x4f\xbc\x32\x21\xa5\xdb"
"\x94\xfa\xe9\x5a\xe7\x12\x1a\x47",
.ilen = 76,
.assoc = "\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xab\xad\xda\xd2",
.alen = 20,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39",
.rlen = 60,
}, {
.key = zeroed_string,
.klen = 24,
.input = "\x98\xe7\x24\x7c\x07\xf0\xfe\x41"
"\x1c\x26\x7e\x43\x84\xb0\xf6\x00"
"\x2f\xf5\x8d\x80\x03\x39\x27\xab"
"\x8e\xf4\xd4\x58\x75\x14\xf0\xfb",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c",
.klen = 24,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x39\x80\xca\x0b\x3c\x00\xe8\x41"
"\xeb\x06\xfa\xc4\x87\x2a\x27\x57"
"\x85\x9e\x1c\xea\xa6\xef\xd9\x84"
"\x62\x85\x93\xb4\x0c\xa1\xe1\x9c"
"\x7d\x77\x3d\x00\xc1\x44\xc5\x25"
"\xac\x61\x9d\x18\xc8\x4a\x3f\x47"
"\x18\xe2\x44\x8b\x2f\xe3\x24\xd9"
"\xcc\xda\x27\x10\xac\xad\xe2\x56"
"\x99\x24\xa7\xc8\x58\x73\x36\xbf"
"\xb1\x18\x02\x4d\xb8\x67\x4a\x14",
.ilen = 80,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39\x1a\xaf\xd2\x55",
.rlen = 64,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\xfe\xff\xe9\x92\x86\x65\x73\x1c",
.klen = 24,
.iv = "\xca\xfe\xba\xbe\xfa\xce\xdb\xad"
"\xde\xca\xf8\x88",
.input = "\x39\x80\xca\x0b\x3c\x00\xe8\x41"
"\xeb\x06\xfa\xc4\x87\x2a\x27\x57"
"\x85\x9e\x1c\xea\xa6\xef\xd9\x84"
"\x62\x85\x93\xb4\x0c\xa1\xe1\x9c"
"\x7d\x77\x3d\x00\xc1\x44\xc5\x25"
"\xac\x61\x9d\x18\xc8\x4a\x3f\x47"
"\x18\xe2\x44\x8b\x2f\xe3\x24\xd9"
"\xcc\xda\x27\x10"
"\x25\x19\x49\x8e\x80\xf1\x47\x8f"
"\x37\xba\x55\xbd\x6d\x27\x61\x8c",
.ilen = 76,
.assoc = "\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xfe\xed\xfa\xce\xde\xad\xbe\xef"
"\xab\xad\xda\xd2",
.alen = 20,
.result = "\xd9\x31\x32\x25\xf8\x84\x06\xe5"
"\xa5\x59\x09\xc5\xaf\xf5\x26\x9a"
"\x86\xa7\xa9\x53\x15\x34\xf7\xda"
"\x2e\x4c\x30\x3d\x8a\x31\x8a\x72"
"\x1c\x3c\x0c\x95\x95\x68\x09\x53"
"\x2f\xcf\x0e\x24\x49\xa6\xb5\x25"
"\xb1\x6a\xed\xf5\xaa\x0d\xe6\x57"
"\xba\x63\x7b\x39",
.rlen = 60,
}
};
static const struct aead_testvec aes_gcm_rfc4106_enc_tv_template[] = {
{ /* Generated using Crypto++ */
.key = zeroed_string,
.klen = 20,
.iv = zeroed_string,
.input = zeroed_string,
.ilen = 16,
.assoc = zeroed_string,
.alen = 16,
.result = "\x03\x88\xDA\xCE\x60\xB6\xA3\x92"
"\xF3\x28\xC2\xB9\x71\xB2\xFE\x78"
"\x97\xFE\x4C\x23\x37\x42\x01\xE0"
"\x81\x9F\x8D\xC5\xD7\x41\xA0\x1B",
.rlen = 32,
},{
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = zeroed_string,
.ilen = 16,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xC0\x0D\x8B\x42\x0F\x8F\x34\x18"
"\x88\xB1\xC5\xBC\xC5\xB6\xD6\x28"
"\x6A\x9D\xDF\x11\x5E\xFE\x5E\x9D"
"\x2F\x70\x44\x92\xF7\xF2\xE3\xEF",
.rlen = 32,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = zeroed_string,
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 16,
.assoc = zeroed_string,
.alen = 16,
.result = "\x4B\xB1\xB5\xE3\x25\x71\x70\xDE"
"\x7F\xC9\x9C\xA5\x14\x19\xF2\xAC"
"\x0B\x8F\x88\x69\x17\xE6\xB4\x3C"
"\xB1\x68\xFD\x14\x52\x64\x61\xB2",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = zeroed_string,
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 16,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.result = "\x4B\xB1\xB5\xE3\x25\x71\x70\xDE"
"\x7F\xC9\x9C\xA5\x14\x19\xF2\xAC"
"\x90\x92\xB7\xE3\x5F\xA3\x9A\x63"
"\x7E\xD7\x1F\xD8\xD3\x7C\x4B\xF5",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 16,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xC1\x0C\x8A\x43\x0E\x8E\x35\x19"
"\x89\xB0\xC4\xBD\xC4\xB7\xD7\x29"
"\x64\x50\xF9\x32\x13\xFB\x74\x61"
"\xF4\xED\x52\xD3\xC5\x10\x55\x3C",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 64,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xC1\x0C\x8A\x43\x0E\x8E\x35\x19"
"\x89\xB0\xC4\xBD\xC4\xB7\xD7\x29"
"\x98\x14\xA1\x42\x37\x80\xFD\x90"
"\x68\x12\x01\xA8\x91\x89\xB9\x83"
"\x5B\x11\x77\x12\x9B\xFF\x24\x89"
"\x94\x5F\x18\x12\xBA\x27\x09\x39"
"\x99\x96\x76\x42\x15\x1C\xCD\xCB"
"\xDC\xD3\xDA\x65\x73\xAF\x80\xCD"
"\xD2\xB6\xC2\x4A\x76\xC2\x92\x85"
"\xBD\xCF\x62\x98\x58\x14\xE5\xBD",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x45\x67\x89\xab\xcd\xef",
.input = "\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff",
.ilen = 192,
.assoc = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\x00\x00\x45\x67"
"\x89\xab\xcd\xef",
.alen = 20,
.result = "\xC1\x76\x33\x85\xE2\x9B\x5F\xDE"
"\xDE\x89\x3D\x42\xE7\xC9\x69\x8A"
"\x44\x6D\xC3\x88\x46\x2E\xC2\x01"
"\x5E\xF6\x0C\x39\xF0\xC4\xA5\x82"
"\xCD\xE8\x31\xCC\x0A\x4C\xE4\x44"
"\x41\xA9\x82\x6F\x22\xA1\x23\x1A"
"\xA8\xE3\x16\xFD\x31\x5C\x27\x31"
"\xF1\x7F\x01\x63\xA3\xAF\x70\xA1"
"\xCF\x07\x57\x41\x67\xD0\xC4\x42"
"\xDB\x18\xC6\x4C\x4C\xE0\x3D\x9F"
"\x05\x07\xFB\x13\x7D\x4A\xCA\x5B"
"\xF0\xBF\x64\x7E\x05\xB1\x72\xEE"
"\x7C\x3B\xD4\xCD\x14\x03\xB2\x2C"
"\xD3\xA9\xEE\xFA\x17\xFC\x9C\xDF"
"\xC7\x75\x40\xFF\xAE\xAD\x1E\x59"
"\x2F\x30\x24\xFB\xAD\x6B\x10\xFA"
"\x6C\x9F\x5B\xE7\x25\xD5\xD0\x25"
"\xAC\x4A\x4B\xDA\xFC\x7A\x85\x1B"
"\x7E\x13\x06\x82\x08\x17\xA4\x35"
"\xEC\xC5\x8D\x63\x96\x81\x0A\x8F"
"\xA3\x05\x38\x95\x20\x1A\x47\x04"
"\x6F\x6D\xDA\x8F\xEF\xC1\x76\x35"
"\x6B\xC7\x4D\x0F\x94\x12\xCA\x3E"
"\x2E\xD5\x03\x2E\x86\x7E\xAA\x3B"
"\x37\x08\x1C\xCF\xBA\x5D\x71\x46"
"\x80\x72\xB0\x4C\x82\x0D\x60\x3C",
.rlen = 208,
}, { /* From draft-mcgrew-gcm-test-01 */
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x2E\x44\x3B\x68",
.klen = 20,
.iv = "\x49\x56\xED\x7E\x3B\x24\x4C\xFE",
.input = "\x45\x00\x00\x48\x69\x9A\x00\x00"
"\x80\x11\x4D\xB7\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x9B\xF1\x56"
"\x38\xD3\x01\x00\x00\x01\x00\x00"
"\x00\x00\x00\x00\x04\x5F\x73\x69"
"\x70\x04\x5F\x75\x64\x70\x03\x73"
"\x69\x70\x09\x63\x79\x62\x65\x72"
"\x63\x69\x74\x79\x02\x64\x6B\x00"
"\x00\x21\x00\x01\x01\x02\x02\x01",
.ilen = 72,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x00\x49\x56\xED\x7E"
"\x3B\x24\x4C\xFE",
.alen = 20,
.result = "\xFE\xCF\x53\x7E\x72\x9D\x5B\x07"
"\xDC\x30\xDF\x52\x8D\xD2\x2B\x76"
"\x8D\x1B\x98\x73\x66\x96\xA6\xFD"
"\x34\x85\x09\xFA\x13\xCE\xAC\x34"
"\xCF\xA2\x43\x6F\x14\xA3\xF3\xCF"
"\x65\x92\x5B\xF1\xF4\xA1\x3C\x5D"
"\x15\xB2\x1E\x18\x84\xF5\xFF\x62"
"\x47\xAE\xAB\xB7\x86\xB9\x3B\xCE"
"\x61\xBC\x17\xD7\x68\xFD\x97\x32"
"\x45\x90\x18\x14\x8F\x6C\xBE\x72"
"\x2F\xD0\x47\x96\x56\x2D\xFD\xB4",
.rlen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xCA\xFE\xBA\xBE",
.klen = 20,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.input = "\x45\x00\x00\x3E\x69\x8F\x00\x00"
"\x80\x11\x4D\xCC\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x98\x00\x35"
"\x00\x2A\x23\x43\xB2\xD0\x01\x00"
"\x00\x01\x00\x00\x00\x00\x00\x00"
"\x03\x73\x69\x70\x09\x63\x79\x62"
"\x65\x72\x63\x69\x74\x79\x02\x64"
"\x6B\x00\x00\x01\x00\x01\x00\x01",
.ilen = 64,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xDE\xB2\x2C\xD9\xB0\x7C\x72\xC1"
"\x6E\x3A\x65\xBE\xEB\x8D\xF3\x04"
"\xA5\xA5\x89\x7D\x33\xAE\x53\x0F"
"\x1B\xA7\x6D\x5D\x11\x4D\x2A\x5C"
"\x3D\xE8\x18\x27\xC1\x0E\x9A\x4F"
"\x51\x33\x0D\x0E\xEC\x41\x66\x42"
"\xCF\xBB\x85\xA5\xB4\x7E\x48\xA4"
"\xEC\x3B\x9B\xA9\x5D\x91\x8B\xD1"
"\x83\xB7\x0D\x3A\xA8\xBC\x6E\xE4"
"\xC3\x09\xE9\xD8\x5A\x41\xAD\x4A",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x11\x22\x33\x44",
.klen = 36,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.input = "\x45\x00\x00\x30\x69\xA6\x40\x00"
"\x80\x06\x26\x90\xC0\xA8\x01\x02"
"\x93\x89\x15\x5E\x0A\x9E\x00\x8B"
"\x2D\xC5\x7E\xE0\x00\x00\x00\x00"
"\x70\x02\x40\x00\x20\xBF\x00\x00"
"\x02\x04\x05\xB4\x01\x01\x04\x02"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x4A\x2C\xBF\xE3\x00\x00\x00\x02"
"\x01\x02\x03\x04\x05\x06\x07\x08",
.alen = 16,
.result = "\xFF\x42\x5C\x9B\x72\x45\x99\xDF"
"\x7A\x3B\xCD\x51\x01\x94\xE0\x0D"
"\x6A\x78\x10\x7F\x1B\x0B\x1C\xBF"
"\x06\xEF\xAE\x9D\x65\xA5\xD7\x63"
"\x74\x8A\x63\x79\x85\x77\x1D\x34"
"\x7F\x05\x45\x65\x9F\x14\xE9\x9D"
"\xEF\x84\x2D\x8E\xB3\x35\xF4\xEE"
"\xCF\xDB\xF8\x31\x82\x4B\x4C\x49"
"\x15\x95\x6C\x96",
.rlen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.input = "\x45\x00\x00\x3C\x99\xC5\x00\x00"
"\x80\x01\xCB\x7A\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x07\x5C"
"\x02\x00\x44\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.ilen = 64,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.result = "\x46\x88\xDA\xF2\xF9\x73\xA3\x92"
"\x73\x29\x09\xC3\x31\xD5\x6D\x60"
"\xF6\x94\xAB\xAA\x41\x4B\x5E\x7F"
"\xF5\xFD\xCD\xFF\xF5\xE9\xA2\x84"
"\x45\x64\x76\x49\x27\x19\xFF\xB6"
"\x4D\xE7\xD9\xDC\xA1\xE1\xD8\x94"
"\xBC\x3B\xD5\x78\x73\xED\x4D\x18"
"\x1D\x19\xD4\xD5\xC8\xC1\x8A\xF3"
"\xF8\x21\xD4\x96\xEE\xB0\x96\xE9"
"\x8A\xD2\xB6\x9E\x47\x99\xC7\x1D",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x3C\x99\xC3\x00\x00"
"\x80\x01\xCB\x7C\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x08\x5C"
"\x02\x00\x43\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.ilen = 64,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\xFB\xA2\xCA\xA4\x85\x3C\xF9\xF0"
"\xF2\x2C\xB1\x0D\x86\xDD\x83\xB0"
"\xFE\xC7\x56\x91\xCF\x1A\x04\xB0"
"\x0D\x11\x38\xEC\x9C\x35\x79\x17"
"\x65\xAC\xBD\x87\x01\xAD\x79\x84"
"\x5B\xF9\xFE\x3F\xBA\x48\x7B\xC9"
"\x17\x55\xE6\x66\x2B\x4C\x8D\x0D"
"\x1F\x5E\x22\x73\x95\x30\x32\x0A"
"\xE0\xD7\x31\xCC\x97\x8E\xCA\xFA"
"\xEA\xE8\x8F\x00\xE8\x0D\x6E\x48",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x1C\x42\xA2\x00\x00"
"\x80\x01\x44\x1F\x40\x67\x93\xB6"
"\xE0\x00\x00\x02\x0A\x00\xF5\xFF"
"\x01\x02\x02\x01",
.ilen = 28,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\xFB\xA2\xCA\x84\x5E\x5D\xF9\xF0"
"\xF2\x2C\x3E\x6E\x86\xDD\x83\x1E"
"\x1F\xC6\x57\x92\xCD\x1A\xF9\x13"
"\x0E\x13\x79\xED\x36\x9F\x07\x1F"
"\x35\xE0\x34\xBE\x95\xF1\x12\xE4"
"\xE7\xD0\x5D\x35",
.rlen = 44,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\xCA\xFE\xBA\xBE",
.klen = 28,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.input = "\x45\x00\x00\x28\xA4\xAD\x40\x00"
"\x40\x06\x78\x80\x0A\x01\x03\x8F"
"\x0A\x01\x06\x12\x80\x23\x06\xB8"
"\xCB\x71\x26\x02\xDD\x6B\xB0\x3E"
"\x50\x10\x16\xD0\x75\x68\x00\x01",
.ilen = 40,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xA5\xB1\xF8\x06\x60\x29\xAE\xA4"
"\x0E\x59\x8B\x81\x22\xDE\x02\x42"
"\x09\x38\xB3\xAB\x33\xF8\x28\xE6"
"\x87\xB8\x85\x8B\x5B\xFB\xDB\xD0"
"\x31\x5B\x27\x45\x21\x44\xCC\x77"
"\x95\x45\x7B\x96\x52\x03\x7F\x53"
"\x18\x02\x7B\x5B\x4C\xD7\xA6\x36",
.rlen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8\x88",
.klen = 20,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.input = "\x45\x00\x00\x49\x33\xBA\x00\x00"
"\x7F\x11\x91\x06\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xDD\x7B\x80\x03\x02\xD5"
"\x00\x00\x4E\x20\x00\x1E\x8C\x18"
"\xD7\x5B\x81\xDC\x91\xBA\xA0\x47"
"\x6B\x91\xB9\x24\xB2\x80\x38\x9D"
"\x92\xC9\x63\xBA\xC0\x46\xEC\x95"
"\x9B\x62\x66\xC0\x47\x22\xB1\x49"
"\x23\x01\x01\x01",
.ilen = 76,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.result = "\x18\xA6\xFD\x42\xF7\x2C\xBF\x4A"
"\xB2\xA2\xEA\x90\x1F\x73\xD8\x14"
"\xE3\xE7\xF2\x43\xD9\x54\x12\xE1"
"\xC3\x49\xC1\xD2\xFB\xEC\x16\x8F"
"\x91\x90\xFE\xEB\xAF\x2C\xB0\x19"
"\x84\xE6\x58\x63\x96\x5D\x74\x72"
"\xB7\x9D\xA3\x45\xE0\xE7\x80\x19"
"\x1F\x0D\x2F\x0E\x0F\x49\x6C\x22"
"\x6F\x21\x27\xB2\x7D\xB3\x57\x24"
"\xE7\x84\x5D\x68\x65\x1F\x57\xE6"
"\x5F\x35\x4F\x75\xFF\x17\x01\x57"
"\x69\x62\x34\x36",
.rlen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C\x74",
.klen = 36,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.input = "\x45\x08\x00\x28\x73\x2C\x00\x00"
"\x40\x06\xE9\xF9\x0A\x01\x06\x12"
"\x0A\x01\x03\x8F\x06\xB8\x80\x23"
"\xDD\x6B\xAF\xBE\xCB\x71\x26\x02"
"\x50\x10\x1F\x64\x6D\x54\x00\x01",
.ilen = 40,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.result = "\xF2\xD6\x9E\xCD\xBD\x5A\x0D\x5B"
"\x8D\x5E\xF3\x8B\xAD\x4D\xA5\x8D"
"\x1F\x27\x8F\xDE\x98\xEF\x67\x54"
"\x9D\x52\x4A\x30\x18\xD9\xA5\x7F"
"\xF4\xD3\xA3\x1C\xE6\x73\x11\x9E"
"\x45\x16\x26\xC2\x41\x57\x71\xE3"
"\xB7\xEE\xBC\xA6\x14\xC8\x9B\x35",
.rlen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x49\x33\x3E\x00\x00"
"\x7F\x11\x91\x82\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xCB\x45\x80\x03\x02\x5B"
"\x00\x00\x01\xE0\x00\x1E\x8C\x18"
"\xD6\x57\x59\xD5\x22\x84\xA0\x35"
"\x2C\x71\x47\x5C\x88\x80\x39\x1C"
"\x76\x4D\x6E\x5E\xE0\x49\x6B\x32"
"\x5A\xE2\x70\xC0\x38\x99\x49\x39"
"\x15\x01\x01\x01",
.ilen = 76,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\xFB\xA2\xCA\xD1\x2F\xC1\xF9\xF0"
"\x0D\x3C\xEB\xF3\x05\x41\x0D\xB8"
"\x3D\x77\x84\xB6\x07\x32\x3D\x22"
"\x0F\x24\xB0\xA9\x7D\x54\x18\x28"
"\x00\xCA\xDB\x0F\x68\xD9\x9E\xF0"
"\xE0\xC0\xC8\x9A\xE9\xBE\xA8\x88"
"\x4E\x52\xD6\x5B\xC1\xAF\xD0\x74"
"\x0F\x74\x24\x44\x74\x7B\x5B\x39"
"\xAB\x53\x31\x63\xAA\xD4\x55\x0E"
"\xE5\x16\x09\x75\xCD\xB6\x08\xC5"
"\x76\x91\x89\x60\x97\x63\xB8\xE1"
"\x8C\xAA\x81\xE2",
.rlen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C\x74",
.klen = 36,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.input = "\x63\x69\x73\x63\x6F\x01\x72\x75"
"\x6C\x65\x73\x01\x74\x68\x65\x01"
"\x6E\x65\x74\x77\x65\x01\x64\x65"
"\x66\x69\x6E\x65\x01\x74\x68\x65"
"\x74\x65\x63\x68\x6E\x6F\x6C\x6F"
"\x67\x69\x65\x73\x01\x74\x68\x61"
"\x74\x77\x69\x6C\x6C\x01\x64\x65"
"\x66\x69\x6E\x65\x74\x6F\x6D\x6F"
"\x72\x72\x6F\x77\x01\x02\x02\x01",
.ilen = 72,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.result = "\xD4\xB7\xED\x86\xA1\x77\x7F\x2E"
"\xA1\x3D\x69\x73\xD3\x24\xC6\x9E"
"\x7B\x43\xF8\x26\xFB\x56\x83\x12"
"\x26\x50\x8B\xEB\xD2\xDC\xEB\x18"
"\xD0\xA6\xDF\x10\xE5\x48\x7D\xF0"
"\x74\x11\x3E\x14\xC6\x41\x02\x4E"
"\x3E\x67\x73\xD9\x1A\x62\xEE\x42"
"\x9B\x04\x3A\x10\xE3\xEF\xE6\xB0"
"\x12\xA4\x93\x63\x41\x23\x64\xF8"
"\xC0\xCA\xC5\x87\xF2\x49\xE5\x6B"
"\x11\xE2\x4F\x30\xE4\x4C\xCC\x76",
.rlen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7D\x77\x3D\x00\xC1\x44\xC5\x25"
"\xAC\x61\x9D\x18\xC8\x4A\x3F\x47"
"\xD9\x66\x42\x67",
.klen = 20,
.iv = "\x43\x45\x7E\x91\x82\x44\x3B\xC6",
.input = "\x01\x02\x02\x01",
.ilen = 4,
.assoc = "\x33\x54\x67\xAE\xFF\xFF\xFF\xFF"
"\x43\x45\x7E\x91\x82\x44\x3B\xC6",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x43\x7F\x86\x6B\xCB\x3F\x69\x9F"
"\xE9\xB0\x82\x2B\xAC\x96\x1C\x45"
"\x04\xBE\xF2\x70",
.rlen = 20,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8\x88",
.klen = 20,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.input = "\x74\x6F\x01\x62\x65\x01\x6F\x72"
"\x01\x6E\x6F\x74\x01\x74\x6F\x01"
"\x62\x65\x00\x01",
.ilen = 20,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.result = "\x29\xC9\xFC\x69\xA1\x97\xD0\x38"
"\xCC\xDD\x14\xE2\xDD\xFC\xAA\x05"
"\x43\x33\x21\x64\x41\x25\x03\x52"
"\x43\x03\xED\x3C\x6C\x5F\x28\x38"
"\x43\xAF\x8C\x3E",
.rlen = 36,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6C\x65\x67\x61\x6C\x69\x7A\x65"
"\x6D\x61\x72\x69\x6A\x75\x61\x6E"
"\x61\x61\x6E\x64\x64\x6F\x69\x74"
"\x62\x65\x66\x6F\x72\x65\x69\x61"
"\x74\x75\x72\x6E",
.klen = 36,
.iv = "\x33\x30\x21\x69\x67\x65\x74\x6D",
.input = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x79\x6B\x69\x63\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\x33\x30\x21\x69"
"\x67\x65\x74\x6D",
.alen = 20,
.result = "\xF9\x7A\xB2\xAA\x35\x6D\x8E\xDC"
"\xE1\x76\x44\xAC\x8C\x78\xE2\x5D"
"\xD2\x4D\xED\xBB\x29\xEB\xF1\xB6"
"\x4A\x27\x4B\x39\xB4\x9C\x3A\x86"
"\x4C\xD3\xD7\x8C\xA4\xAE\x68\xA3"
"\x2B\x42\x45\x8F\xB5\x7D\xBE\x82"
"\x1D\xCC\x63\xB9\xD0\x93\x7B\xA2"
"\x94\x5F\x66\x93\x68\x66\x1A\x32"
"\x9F\xB4\xC0\x53",
.rlen = 68,
}, {
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x3F\x7E\xF6\x42\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\xFB\xA2\xCA\xA8\xC6\xC5\xF9\xF0"
"\xF2\x2C\xA5\x4A\x06\x12\x10\xAD"
"\x3F\x6E\x57\x91\xCF\x1A\xCA\x21"
"\x0D\x11\x7C\xEC\x9C\x35\x79\x17"
"\x65\xAC\xBD\x87\x01\xAD\x79\x84"
"\x5B\xF9\xFE\x3F\xBA\x48\x7B\xC9"
"\x63\x21\x93\x06\x84\xEE\xCA\xDB"
"\x56\x91\x25\x46\xE7\xA9\x5C\x97"
"\x40\xD7\xCB\x05",
.rlen = 68,
}, {
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x22\x43\x3C\x64",
.klen = 20,
.iv = "\x48\x55\xEC\x7D\x3A\x23\x4B\xFD",
.input = "\x08\x00\xC6\xCD\x02\x00\x07\x00"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6A\x6B\x6C\x6D\x6E\x6F\x70"
"\x71\x72\x73\x74\x01\x02\x02\x01",
.ilen = 32,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x07\x48\x55\xEC\x7D"
"\x3A\x23\x4B\xFD",
.alen = 20,
.result = "\x74\x75\x2E\x8A\xEB\x5D\x87\x3C"
"\xD7\xC0\xF4\xAC\xC3\x6C\x4B\xFF"
"\x84\xB7\xD7\xB9\x8F\x0C\xA8\xB6"
"\xAC\xDA\x68\x94\xBC\x61\x90\x69"
"\xEF\x9C\xBC\x28\xFE\x1B\x56\xA7"
"\xC4\xE0\xD5\x8C\x86\xCD\x2B\xC0",
.rlen = 48,
}
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aes_gcm_rfc4106_dec_tv_template[] = {
{ /* Generated using Crypto++ */
.key = zeroed_string,
.klen = 20,
.iv = zeroed_string,
.input = "\x03\x88\xDA\xCE\x60\xB6\xA3\x92"
"\xF3\x28\xC2\xB9\x71\xB2\xFE\x78"
"\x97\xFE\x4C\x23\x37\x42\x01\xE0"
"\x81\x9F\x8D\xC5\xD7\x41\xA0\x1B",
.ilen = 32,
.assoc = zeroed_string,
.alen = 16,
.result = zeroed_string,
.rlen = 16,
},{
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\xC0\x0D\x8B\x42\x0F\x8F\x34\x18"
"\x88\xB1\xC5\xBC\xC5\xB6\xD6\x28"
"\x6A\x9D\xDF\x11\x5E\xFE\x5E\x9D"
"\x2F\x70\x44\x92\xF7\xF2\xE3\xEF",
.ilen = 32,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = zeroed_string,
.rlen = 16,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = zeroed_string,
.input = "\x4B\xB1\xB5\xE3\x25\x71\x70\xDE"
"\x7F\xC9\x9C\xA5\x14\x19\xF2\xAC"
"\x0B\x8F\x88\x69\x17\xE6\xB4\x3C"
"\xB1\x68\xFD\x14\x52\x64\x61\xB2",
.ilen = 32,
.assoc = zeroed_string,
.alen = 16,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
}, {
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = zeroed_string,
.input = "\x4B\xB1\xB5\xE3\x25\x71\x70\xDE"
"\x7F\xC9\x9C\xA5\x14\x19\xF2\xAC"
"\x90\x92\xB7\xE3\x5F\xA3\x9A\x63"
"\x7E\xD7\x1F\xD8\xD3\x7C\x4B\xF5",
.ilen = 32,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\xC1\x0C\x8A\x43\x0E\x8E\x35\x19"
"\x89\xB0\xC4\xBD\xC4\xB7\xD7\x29"
"\x64\x50\xF9\x32\x13\xFB\x74\x61"
"\xF4\xED\x52\xD3\xC5\x10\x55\x3C",
.ilen = 32,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\xC1\x0C\x8A\x43\x0E\x8E\x35\x19"
"\x89\xB0\xC4\xBD\xC4\xB7\xD7\x29"
"\x98\x14\xA1\x42\x37\x80\xFD\x90"
"\x68\x12\x01\xA8\x91\x89\xB9\x83"
"\x5B\x11\x77\x12\x9B\xFF\x24\x89"
"\x94\x5F\x18\x12\xBA\x27\x09\x39"
"\x99\x96\x76\x42\x15\x1C\xCD\xCB"
"\xDC\xD3\xDA\x65\x73\xAF\x80\xCD"
"\xD2\xB6\xC2\x4A\x76\xC2\x92\x85"
"\xBD\xCF\x62\x98\x58\x14\xE5\xBD",
.ilen = 80,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 64,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x45\x67\x89\xab\xcd\xef",
.input = "\xC1\x76\x33\x85\xE2\x9B\x5F\xDE"
"\xDE\x89\x3D\x42\xE7\xC9\x69\x8A"
"\x44\x6D\xC3\x88\x46\x2E\xC2\x01"
"\x5E\xF6\x0C\x39\xF0\xC4\xA5\x82"
"\xCD\xE8\x31\xCC\x0A\x4C\xE4\x44"
"\x41\xA9\x82\x6F\x22\xA1\x23\x1A"
"\xA8\xE3\x16\xFD\x31\x5C\x27\x31"
"\xF1\x7F\x01\x63\xA3\xAF\x70\xA1"
"\xCF\x07\x57\x41\x67\xD0\xC4\x42"
"\xDB\x18\xC6\x4C\x4C\xE0\x3D\x9F"
"\x05\x07\xFB\x13\x7D\x4A\xCA\x5B"
"\xF0\xBF\x64\x7E\x05\xB1\x72\xEE"
"\x7C\x3B\xD4\xCD\x14\x03\xB2\x2C"
"\xD3\xA9\xEE\xFA\x17\xFC\x9C\xDF"
"\xC7\x75\x40\xFF\xAE\xAD\x1E\x59"
"\x2F\x30\x24\xFB\xAD\x6B\x10\xFA"
"\x6C\x9F\x5B\xE7\x25\xD5\xD0\x25"
"\xAC\x4A\x4B\xDA\xFC\x7A\x85\x1B"
"\x7E\x13\x06\x82\x08\x17\xA4\x35"
"\xEC\xC5\x8D\x63\x96\x81\x0A\x8F"
"\xA3\x05\x38\x95\x20\x1A\x47\x04"
"\x6F\x6D\xDA\x8F\xEF\xC1\x76\x35"
"\x6B\xC7\x4D\x0F\x94\x12\xCA\x3E"
"\x2E\xD5\x03\x2E\x86\x7E\xAA\x3B"
"\x37\x08\x1C\xCF\xBA\x5D\x71\x46"
"\x80\x72\xB0\x4C\x82\x0D\x60\x3C",
.ilen = 208,
.assoc = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\x00\x00\x45\x67"
"\x89\xab\xcd\xef",
.alen = 20,
.result = "\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff",
.rlen = 192,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x2E\x44\x3B\x68",
.klen = 20,
.iv = "\x49\x56\xED\x7E\x3B\x24\x4C\xFE",
.result = "\x45\x00\x00\x48\x69\x9A\x00\x00"
"\x80\x11\x4D\xB7\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x9B\xF1\x56"
"\x38\xD3\x01\x00\x00\x01\x00\x00"
"\x00\x00\x00\x00\x04\x5F\x73\x69"
"\x70\x04\x5F\x75\x64\x70\x03\x73"
"\x69\x70\x09\x63\x79\x62\x65\x72"
"\x63\x69\x74\x79\x02\x64\x6B\x00"
"\x00\x21\x00\x01\x01\x02\x02\x01",
.rlen = 72,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x00\x49\x56\xED\x7E"
"\x3B\x24\x4C\xFE",
.alen = 20,
.input = "\xFE\xCF\x53\x7E\x72\x9D\x5B\x07"
"\xDC\x30\xDF\x52\x8D\xD2\x2B\x76"
"\x8D\x1B\x98\x73\x66\x96\xA6\xFD"
"\x34\x85\x09\xFA\x13\xCE\xAC\x34"
"\xCF\xA2\x43\x6F\x14\xA3\xF3\xCF"
"\x65\x92\x5B\xF1\xF4\xA1\x3C\x5D"
"\x15\xB2\x1E\x18\x84\xF5\xFF\x62"
"\x47\xAE\xAB\xB7\x86\xB9\x3B\xCE"
"\x61\xBC\x17\xD7\x68\xFD\x97\x32"
"\x45\x90\x18\x14\x8F\x6C\xBE\x72"
"\x2F\xD0\x47\x96\x56\x2D\xFD\xB4",
.ilen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xCA\xFE\xBA\xBE",
.klen = 20,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.result = "\x45\x00\x00\x3E\x69\x8F\x00\x00"
"\x80\x11\x4D\xCC\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x98\x00\x35"
"\x00\x2A\x23\x43\xB2\xD0\x01\x00"
"\x00\x01\x00\x00\x00\x00\x00\x00"
"\x03\x73\x69\x70\x09\x63\x79\x62"
"\x65\x72\x63\x69\x74\x79\x02\x64"
"\x6B\x00\x00\x01\x00\x01\x00\x01",
.rlen = 64,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.alen = 16,
.input = "\xDE\xB2\x2C\xD9\xB0\x7C\x72\xC1"
"\x6E\x3A\x65\xBE\xEB\x8D\xF3\x04"
"\xA5\xA5\x89\x7D\x33\xAE\x53\x0F"
"\x1B\xA7\x6D\x5D\x11\x4D\x2A\x5C"
"\x3D\xE8\x18\x27\xC1\x0E\x9A\x4F"
"\x51\x33\x0D\x0E\xEC\x41\x66\x42"
"\xCF\xBB\x85\xA5\xB4\x7E\x48\xA4"
"\xEC\x3B\x9B\xA9\x5D\x91\x8B\xD1"
"\x83\xB7\x0D\x3A\xA8\xBC\x6E\xE4"
"\xC3\x09\xE9\xD8\x5A\x41\xAD\x4A",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x11\x22\x33\x44",
.klen = 36,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.result = "\x45\x00\x00\x30\x69\xA6\x40\x00"
"\x80\x06\x26\x90\xC0\xA8\x01\x02"
"\x93\x89\x15\x5E\x0A\x9E\x00\x8B"
"\x2D\xC5\x7E\xE0\x00\x00\x00\x00"
"\x70\x02\x40\x00\x20\xBF\x00\x00"
"\x02\x04\x05\xB4\x01\x01\x04\x02"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x4A\x2C\xBF\xE3\x00\x00\x00\x02"
"\x01\x02\x03\x04\x05\x06\x07\x08",
.alen = 16,
.input = "\xFF\x42\x5C\x9B\x72\x45\x99\xDF"
"\x7A\x3B\xCD\x51\x01\x94\xE0\x0D"
"\x6A\x78\x10\x7F\x1B\x0B\x1C\xBF"
"\x06\xEF\xAE\x9D\x65\xA5\xD7\x63"
"\x74\x8A\x63\x79\x85\x77\x1D\x34"
"\x7F\x05\x45\x65\x9F\x14\xE9\x9D"
"\xEF\x84\x2D\x8E\xB3\x35\xF4\xEE"
"\xCF\xDB\xF8\x31\x82\x4B\x4C\x49"
"\x15\x95\x6C\x96",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.klen = 20,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.result = "\x45\x00\x00\x3C\x99\xC5\x00\x00"
"\x80\x01\xCB\x7A\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x07\x5C"
"\x02\x00\x44\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.rlen = 64,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.input = "\x46\x88\xDA\xF2\xF9\x73\xA3\x92"
"\x73\x29\x09\xC3\x31\xD5\x6D\x60"
"\xF6\x94\xAB\xAA\x41\x4B\x5E\x7F"
"\xF5\xFD\xCD\xFF\xF5\xE9\xA2\x84"
"\x45\x64\x76\x49\x27\x19\xFF\xB6"
"\x4D\xE7\xD9\xDC\xA1\xE1\xD8\x94"
"\xBC\x3B\xD5\x78\x73\xED\x4D\x18"
"\x1D\x19\xD4\xD5\xC8\xC1\x8A\xF3"
"\xF8\x21\xD4\x96\xEE\xB0\x96\xE9"
"\x8A\xD2\xB6\x9E\x47\x99\xC7\x1D",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x3C\x99\xC3\x00\x00"
"\x80\x01\xCB\x7C\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x08\x5C"
"\x02\x00\x43\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.rlen = 64,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\xFB\xA2\xCA\xA4\x85\x3C\xF9\xF0"
"\xF2\x2C\xB1\x0D\x86\xDD\x83\xB0"
"\xFE\xC7\x56\x91\xCF\x1A\x04\xB0"
"\x0D\x11\x38\xEC\x9C\x35\x79\x17"
"\x65\xAC\xBD\x87\x01\xAD\x79\x84"
"\x5B\xF9\xFE\x3F\xBA\x48\x7B\xC9"
"\x17\x55\xE6\x66\x2B\x4C\x8D\x0D"
"\x1F\x5E\x22\x73\x95\x30\x32\x0A"
"\xE0\xD7\x31\xCC\x97\x8E\xCA\xFA"
"\xEA\xE8\x8F\x00\xE8\x0D\x6E\x48",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x1C\x42\xA2\x00\x00"
"\x80\x01\x44\x1F\x40\x67\x93\xB6"
"\xE0\x00\x00\x02\x0A\x00\xF5\xFF"
"\x01\x02\x02\x01",
.rlen = 28,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\xFB\xA2\xCA\x84\x5E\x5D\xF9\xF0"
"\xF2\x2C\x3E\x6E\x86\xDD\x83\x1E"
"\x1F\xC6\x57\x92\xCD\x1A\xF9\x13"
"\x0E\x13\x79\xED\x36\x9F\x07\x1F"
"\x35\xE0\x34\xBE\x95\xF1\x12\xE4"
"\xE7\xD0\x5D\x35",
.ilen = 44,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\xCA\xFE\xBA\xBE",
.klen = 28,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.result = "\x45\x00\x00\x28\xA4\xAD\x40\x00"
"\x40\x06\x78\x80\x0A\x01\x03\x8F"
"\x0A\x01\x06\x12\x80\x23\x06\xB8"
"\xCB\x71\x26\x02\xDD\x6B\xB0\x3E"
"\x50\x10\x16\xD0\x75\x68\x00\x01",
.rlen = 40,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.alen = 16,
.input = "\xA5\xB1\xF8\x06\x60\x29\xAE\xA4"
"\x0E\x59\x8B\x81\x22\xDE\x02\x42"
"\x09\x38\xB3\xAB\x33\xF8\x28\xE6"
"\x87\xB8\x85\x8B\x5B\xFB\xDB\xD0"
"\x31\x5B\x27\x45\x21\x44\xCC\x77"
"\x95\x45\x7B\x96\x52\x03\x7F\x53"
"\x18\x02\x7B\x5B\x4C\xD7\xA6\x36",
.ilen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8\x88",
.klen = 20,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.result = "\x45\x00\x00\x49\x33\xBA\x00\x00"
"\x7F\x11\x91\x06\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xDD\x7B\x80\x03\x02\xD5"
"\x00\x00\x4E\x20\x00\x1E\x8C\x18"
"\xD7\x5B\x81\xDC\x91\xBA\xA0\x47"
"\x6B\x91\xB9\x24\xB2\x80\x38\x9D"
"\x92\xC9\x63\xBA\xC0\x46\xEC\x95"
"\x9B\x62\x66\xC0\x47\x22\xB1\x49"
"\x23\x01\x01\x01",
.rlen = 76,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.input = "\x18\xA6\xFD\x42\xF7\x2C\xBF\x4A"
"\xB2\xA2\xEA\x90\x1F\x73\xD8\x14"
"\xE3\xE7\xF2\x43\xD9\x54\x12\xE1"
"\xC3\x49\xC1\xD2\xFB\xEC\x16\x8F"
"\x91\x90\xFE\xEB\xAF\x2C\xB0\x19"
"\x84\xE6\x58\x63\x96\x5D\x74\x72"
"\xB7\x9D\xA3\x45\xE0\xE7\x80\x19"
"\x1F\x0D\x2F\x0E\x0F\x49\x6C\x22"
"\x6F\x21\x27\xB2\x7D\xB3\x57\x24"
"\xE7\x84\x5D\x68\x65\x1F\x57\xE6"
"\x5F\x35\x4F\x75\xFF\x17\x01\x57"
"\x69\x62\x34\x36",
.ilen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C\x74",
.klen = 36,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.result = "\x45\x08\x00\x28\x73\x2C\x00\x00"
"\x40\x06\xE9\xF9\x0A\x01\x06\x12"
"\x0A\x01\x03\x8F\x06\xB8\x80\x23"
"\xDD\x6B\xAF\xBE\xCB\x71\x26\x02"
"\x50\x10\x1F\x64\x6D\x54\x00\x01",
.rlen = 40,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.input = "\xF2\xD6\x9E\xCD\xBD\x5A\x0D\x5B"
"\x8D\x5E\xF3\x8B\xAD\x4D\xA5\x8D"
"\x1F\x27\x8F\xDE\x98\xEF\x67\x54"
"\x9D\x52\x4A\x30\x18\xD9\xA5\x7F"
"\xF4\xD3\xA3\x1C\xE6\x73\x11\x9E"
"\x45\x16\x26\xC2\x41\x57\x71\xE3"
"\xB7\xEE\xBC\xA6\x14\xC8\x9B\x35",
.ilen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x49\x33\x3E\x00\x00"
"\x7F\x11\x91\x82\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xCB\x45\x80\x03\x02\x5B"
"\x00\x00\x01\xE0\x00\x1E\x8C\x18"
"\xD6\x57\x59\xD5\x22\x84\xA0\x35"
"\x2C\x71\x47\x5C\x88\x80\x39\x1C"
"\x76\x4D\x6E\x5E\xE0\x49\x6B\x32"
"\x5A\xE2\x70\xC0\x38\x99\x49\x39"
"\x15\x01\x01\x01",
.rlen = 76,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\xFB\xA2\xCA\xD1\x2F\xC1\xF9\xF0"
"\x0D\x3C\xEB\xF3\x05\x41\x0D\xB8"
"\x3D\x77\x84\xB6\x07\x32\x3D\x22"
"\x0F\x24\xB0\xA9\x7D\x54\x18\x28"
"\x00\xCA\xDB\x0F\x68\xD9\x9E\xF0"
"\xE0\xC0\xC8\x9A\xE9\xBE\xA8\x88"
"\x4E\x52\xD6\x5B\xC1\xAF\xD0\x74"
"\x0F\x74\x24\x44\x74\x7B\x5B\x39"
"\xAB\x53\x31\x63\xAA\xD4\x55\x0E"
"\xE5\x16\x09\x75\xCD\xB6\x08\xC5"
"\x76\x91\x89\x60\x97\x63\xB8\xE1"
"\x8C\xAA\x81\xE2",
.ilen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C\x74",
.klen = 36,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.result = "\x63\x69\x73\x63\x6F\x01\x72\x75"
"\x6C\x65\x73\x01\x74\x68\x65\x01"
"\x6E\x65\x74\x77\x65\x01\x64\x65"
"\x66\x69\x6E\x65\x01\x74\x68\x65"
"\x74\x65\x63\x68\x6E\x6F\x6C\x6F"
"\x67\x69\x65\x73\x01\x74\x68\x61"
"\x74\x77\x69\x6C\x6C\x01\x64\x65"
"\x66\x69\x6E\x65\x74\x6F\x6D\x6F"
"\x72\x72\x6F\x77\x01\x02\x02\x01",
.rlen = 72,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.input = "\xD4\xB7\xED\x86\xA1\x77\x7F\x2E"
"\xA1\x3D\x69\x73\xD3\x24\xC6\x9E"
"\x7B\x43\xF8\x26\xFB\x56\x83\x12"
"\x26\x50\x8B\xEB\xD2\xDC\xEB\x18"
"\xD0\xA6\xDF\x10\xE5\x48\x7D\xF0"
"\x74\x11\x3E\x14\xC6\x41\x02\x4E"
"\x3E\x67\x73\xD9\x1A\x62\xEE\x42"
"\x9B\x04\x3A\x10\xE3\xEF\xE6\xB0"
"\x12\xA4\x93\x63\x41\x23\x64\xF8"
"\xC0\xCA\xC5\x87\xF2\x49\xE5\x6B"
"\x11\xE2\x4F\x30\xE4\x4C\xCC\x76",
.ilen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7D\x77\x3D\x00\xC1\x44\xC5\x25"
"\xAC\x61\x9D\x18\xC8\x4A\x3F\x47"
"\xD9\x66\x42\x67",
.klen = 20,
.iv = "\x43\x45\x7E\x91\x82\x44\x3B\xC6",
.result = "\x01\x02\x02\x01",
.rlen = 4,
.assoc = "\x33\x54\x67\xAE\xFF\xFF\xFF\xFF"
"\x43\x45\x7E\x91\x82\x44\x3B\xC6",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x43\x7F\x86\x6B\xCB\x3F\x69\x9F"
"\xE9\xB0\x82\x2B\xAC\x96\x1C\x45"
"\x04\xBE\xF2\x70",
.ilen = 20,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8\x88",
.klen = 20,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.result = "\x74\x6F\x01\x62\x65\x01\x6F\x72"
"\x01\x6E\x6F\x74\x01\x74\x6F\x01"
"\x62\x65\x00\x01",
.rlen = 20,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.input = "\x29\xC9\xFC\x69\xA1\x97\xD0\x38"
"\xCC\xDD\x14\xE2\xDD\xFC\xAA\x05"
"\x43\x33\x21\x64\x41\x25\x03\x52"
"\x43\x03\xED\x3C\x6C\x5F\x28\x38"
"\x43\xAF\x8C\x3E",
.ilen = 36,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6C\x65\x67\x61\x6C\x69\x7A\x65"
"\x6D\x61\x72\x69\x6A\x75\x61\x6E"
"\x61\x61\x6E\x64\x64\x6F\x69\x74"
"\x62\x65\x66\x6F\x72\x65\x69\x61"
"\x74\x75\x72\x6E",
.klen = 36,
.iv = "\x33\x30\x21\x69\x67\x65\x74\x6D",
.result = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x79\x6B\x69\x63\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\x33\x30\x21\x69"
"\x67\x65\x74\x6D",
.alen = 20,
.input = "\xF9\x7A\xB2\xAA\x35\x6D\x8E\xDC"
"\xE1\x76\x44\xAC\x8C\x78\xE2\x5D"
"\xD2\x4D\xED\xBB\x29\xEB\xF1\xB6"
"\x4A\x27\x4B\x39\xB4\x9C\x3A\x86"
"\x4C\xD3\xD7\x8C\xA4\xAE\x68\xA3"
"\x2B\x42\x45\x8F\xB5\x7D\xBE\x82"
"\x1D\xCC\x63\xB9\xD0\x93\x7B\xA2"
"\x94\x5F\x66\x93\x68\x66\x1A\x32"
"\x9F\xB4\xC0\x53",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E\x43",
.klen = 20,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x3F\x7E\xF6\x42\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\xFB\xA2\xCA\xA8\xC6\xC5\xF9\xF0"
"\xF2\x2C\xA5\x4A\x06\x12\x10\xAD"
"\x3F\x6E\x57\x91\xCF\x1A\xCA\x21"
"\x0D\x11\x7C\xEC\x9C\x35\x79\x17"
"\x65\xAC\xBD\x87\x01\xAD\x79\x84"
"\x5B\xF9\xFE\x3F\xBA\x48\x7B\xC9"
"\x63\x21\x93\x06\x84\xEE\xCA\xDB"
"\x56\x91\x25\x46\xE7\xA9\x5C\x97"
"\x40\xD7\xCB\x05",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x22\x43\x3C\x64",
.klen = 20,
.iv = "\x48\x55\xEC\x7D\x3A\x23\x4B\xFD",
.result = "\x08\x00\xC6\xCD\x02\x00\x07\x00"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6A\x6B\x6C\x6D\x6E\x6F\x70"
"\x71\x72\x73\x74\x01\x02\x02\x01",
.rlen = 32,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x07\x48\x55\xEC\x7D"
"\x3A\x23\x4B\xFD",
.alen = 20,
.input = "\x74\x75\x2E\x8A\xEB\x5D\x87\x3C"
"\xD7\xC0\xF4\xAC\xC3\x6C\x4B\xFF"
"\x84\xB7\xD7\xB9\x8F\x0C\xA8\xB6"
"\xAC\xDA\x68\x94\xBC\x61\x90\x69"
"\xEF\x9C\xBC\x28\xFE\x1B\x56\xA7"
"\xC4\xE0\xD5\x8C\x86\xCD\x2B\xC0",
.ilen = 48,
}
};
static const struct aead_testvec aes_gcm_rfc4543_enc_tv_template[] = {
{ /* From draft-mcgrew-gcm-test-01 */
.key = "\x4c\x80\xcd\xef\xbb\x5d\x10\xda"
"\x90\x6a\xc7\x3c\x36\x13\xa6\x34"
"\x22\x43\x3c\x64",
.klen = 20,
.iv = zeroed_string,
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x07"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.ilen = 52,
.result = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01\xf2\xa9\xa8\x36"
"\xe1\x55\x10\x6a\xa8\xdc\xd6\x18"
"\xe4\x09\x9a\xaa",
.rlen = 68,
}
};
static const struct aead_testvec aes_gcm_rfc4543_dec_tv_template[] = {
{ /* From draft-mcgrew-gcm-test-01 */
.key = "\x4c\x80\xcd\xef\xbb\x5d\x10\xda"
"\x90\x6a\xc7\x3c\x36\x13\xa6\x34"
"\x22\x43\x3c\x64",
.klen = 20,
.iv = zeroed_string,
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x07"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01\xf2\xa9\xa8\x36"
"\xe1\x55\x10\x6a\xa8\xdc\xd6\x18"
"\xe4\x09\x9a\xaa",
.ilen = 68,
.result = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
}, { /* nearly same as previous, but should fail */
.key = "\x4c\x80\xcd\xef\xbb\x5d\x10\xda"
"\x90\x6a\xc7\x3c\x36\x13\xa6\x34"
"\x22\x43\x3c\x64",
.klen = 20,
.iv = zeroed_string,
.assoc = "\x00\x00\x43\x21\x00\x00\x00\x07"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01\xf2\xa9\xa8\x36"
"\xe1\x55\x10\x6a\xa8\xdc\xd6\x18"
"\x00\x00\x00\x00",
.ilen = 68,
.novrfy = 1,
.result = "\x45\x00\x00\x30\xda\x3a\x00\x00"
"\x80\x01\xdf\x3b\xc0\xa8\x00\x05"
"\xc0\xa8\x00\x01\x08\x00\xc6\xcd"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6a\x6b\x6c"
"\x6d\x6e\x6f\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aes_ccm_enc_tv_template[] = {
{ /* From RFC 3610 */
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x03\x02\x01\x00"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07",
.alen = 8,
.input = "\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e",
.ilen = 23,
.result = "\x58\x8c\x97\x9a\x61\xc6\x63\xd2"
"\xf0\x66\xd0\xc2\xc0\xf9\x89\x80"
"\x6d\x5f\x6b\x61\xda\xc3\x84\x17"
"\xe8\xd1\x2c\xfd\xf9\x26\xe0",
.rlen = 31,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x07\x06\x05\x04"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b",
.alen = 12,
.input = "\x0c\x0d\x0e\x0f\x10\x11\x12\x13"
"\x14\x15\x16\x17\x18\x19\x1a\x1b"
"\x1c\x1d\x1e\x1f",
.ilen = 20,
.result = "\xdc\xf1\xfb\x7b\x5d\x9e\x23\xfb"
"\x9d\x4e\x13\x12\x53\x65\x8a\xd8"
"\x6e\xbd\xca\x3e\x51\xe8\x3f\x07"
"\x7d\x9c\x2d\x93",
.rlen = 28,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x0b\x0a\x09\x08"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07",
.alen = 8,
.input = "\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20",
.ilen = 25,
.result = "\x82\x53\x1a\x60\xcc\x24\x94\x5a"
"\x4b\x82\x79\x18\x1a\xb5\xc8\x4d"
"\xf2\x1c\xe7\xf9\xb7\x3f\x42\xe1"
"\x97\xea\x9c\x07\xe5\x6b\x5e\xb1"
"\x7e\x5f\x4e",
.rlen = 35,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x0c\x0b\x0a\x09"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b",
.alen = 12,
.input = "\x0c\x0d\x0e\x0f\x10\x11\x12\x13"
"\x14\x15\x16\x17\x18\x19\x1a\x1b"
"\x1c\x1d\x1e",
.ilen = 19,
.result = "\x07\x34\x25\x94\x15\x77\x85\x15"
"\x2b\x07\x40\x98\x33\x0a\xbb\x14"
"\x1b\x94\x7b\x56\x6a\xa9\x40\x6b"
"\x4d\x99\x99\x88\xdd",
.rlen = 29,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x33\x56\x8e\xf7\xb2\x63"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\x63\x01\x8f\x76\xdc\x8a\x1b\xcb",
.alen = 8,
.input = "\x90\x20\xea\x6f\x91\xbd\xd8\x5a"
"\xfa\x00\x39\xba\x4b\xaf\xf9\xbf"
"\xb7\x9c\x70\x28\x94\x9c\xd0\xec",
.ilen = 24,
.result = "\x4c\xcb\x1e\x7c\xa9\x81\xbe\xfa"
"\xa0\x72\x6c\x55\xd3\x78\x06\x12"
"\x98\xc8\x5c\x92\x81\x4a\xbc\x33"
"\xc5\x2e\xe8\x1d\x7d\x77\xc0\x8a",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\xd5\x60\x91\x2d\x3f\x70"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\xcd\x90\x44\xd2\xb7\x1f\xdb\x81"
"\x20\xea\x60\xc0",
.alen = 12,
.input = "\x64\x35\xac\xba\xfb\x11\xa8\x2e"
"\x2f\x07\x1d\x7c\xa4\xa5\xeb\xd9"
"\x3a\x80\x3b\xa8\x7f",
.ilen = 21,
.result = "\x00\x97\x69\xec\xab\xdf\x48\x62"
"\x55\x94\xc5\x92\x51\xe6\x03\x57"
"\x22\x67\x5e\x04\xc8\x47\x09\x9e"
"\x5a\xe0\x70\x45\x51",
.rlen = 29,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x42\xff\xf8\xf1\x95\x1c"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\xd8\x5b\xc7\xe6\x9f\x94\x4f\xb8",
.alen = 8,
.input = "\x8a\x19\xb9\x50\xbc\xf7\x1a\x01"
"\x8e\x5e\x67\x01\xc9\x17\x87\x65"
"\x98\x09\xd6\x7d\xbe\xdd\x18",
.ilen = 23,
.result = "\xbc\x21\x8d\xaa\x94\x74\x27\xb6"
"\xdb\x38\x6a\x99\xac\x1a\xef\x23"
"\xad\xe0\xb5\x29\x39\xcb\x6a\x63"
"\x7c\xf9\xbe\xc2\x40\x88\x97\xc6"
"\xba",
.rlen = 33,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/* This is taken from FIPS CAVS. */
.key = "\x83\xac\x54\x66\xc2\xeb\xe5\x05"
"\x2e\x01\xd1\xfc\x5d\x82\x66\x2e",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x03\x96\xac\x59\x30\x07\xa1\xe2\xa2\xc7\x55\x24\0\0\0\0",
.alen = 0,
.input = "\x19\xc8\x81\xf6\xe9\x86\xff\x93"
"\x0b\x78\x67\xe5\xbb\xb7\xfc\x6e"
"\x83\x77\xb3\xa6\x0c\x8c\x9f\x9c"
"\x35\x2e\xad\xe0\x62\xf9\x91\xa1",
.ilen = 32,
.result = "\xab\x6f\xe1\x69\x1d\x19\x99\xa8"
"\x92\xa0\xc4\x6f\x7e\xe2\x8b\xb1"
"\x70\xbb\x8c\xa6\x4c\x6e\x97\x8a"
"\x57\x2b\xbe\x5d\x98\xa6\xb1\x32"
"\xda\x24\xea\xd9\xa1\x39\x98\xfd"
"\xa4\xbe\xd9\xf2\x1a\x6d\x22\xa8",
.rlen = 48,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x1e\x2c\x7e\x01\x41\x9a\xef\xc0"
"\x0d\x58\x96\x6e\x5c\xa2\x4b\xd3",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x03\x4f\xa3\x19\xd3\x01\x5a\xd8"
"\x30\x60\x15\x56\x00\x00\x00\x00",
.assoc = "\xda\xe6\x28\x9c\x45\x2d\xfd\x63"
"\x5e\xda\x4c\xb6\xe6\xfc\xf9\xb7"
"\x0c\x56\xcb\xe4\xe0\x05\x7a\xe1"
"\x0a\x63\x09\x78\xbc\x2c\x55\xde",
.alen = 32,
.input = "\x87\xa3\x36\xfd\x96\xb3\x93\x78"
"\xa9\x28\x63\xba\x12\xa3\x14\x85"
"\x57\x1e\x06\xc9\x7b\x21\xef\x76"
"\x7f\x38\x7e\x8e\x29\xa4\x3e\x7e",
.ilen = 32,
.result = "\x8a\x1e\x11\xf0\x02\x6b\xe2\x19"
"\xfc\x70\xc4\x6d\x8e\xb7\x99\xab"
"\xc5\x4b\xa2\xac\xd3\xf3\x48\xff"
"\x3b\xb5\xce\x53\xef\xde\xbb\x02"
"\xa9\x86\x15\x6c\x13\xfe\xda\x0a"
"\x22\xb8\x29\x3d\xd8\x39\x9a\x23",
.rlen = 48,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf4\x6b\xc2\x75\x62\xfe\xb4\xe1"
"\xa3\xf0\xff\xdd\x4e\x4b\x12\x75"
"\x53\x14\x73\x66\x8d\x88\xf6\x80",
.klen = 24,
.iv = "\x03\xa0\x20\x35\x26\xf2\x21\x8d"
"\x50\x20\xda\xe2\x00\x00\x00\x00",
.assoc = "\x5b\x9e\x13\x67\x02\x5e\xef\xc1"
"\x6c\xf9\xd7\x1e\x52\x8f\x7a\x47"
"\xe9\xd4\xcf\x20\x14\x6e\xf0\x2d"
"\xd8\x9e\x2b\x56\x10\x23\x56\xe7",
.alen = 32,
.result = "\x36\xea\x7a\x70\x08\xdc\x6a\xbc"
"\xad\x0c\x7a\x63\xf6\x61\xfd\x9b",
.rlen = 16,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x56\xdf\x5c\x8f\x26\x3f\x0e\x42"
"\xef\x7a\xd3\xce\xfc\x84\x60\x62"
"\xca\xb4\x40\xaf\x5f\xc9\xc9\x01",
.klen = 24,
.iv = "\x03\xd6\x3c\x8c\x86\x84\xb6\xcd"
"\xef\x09\x2e\x94\x00\x00\x00\x00",
.assoc = "\x02\x65\x78\x3c\xe9\x21\x30\x91"
"\xb1\xb9\xda\x76\x9a\x78\x6d\x95"
"\xf2\x88\x32\xa3\xf2\x50\xcb\x4c"
"\xe3\x00\x73\x69\x84\x69\x87\x79",
.alen = 32,
.input = "\x9f\xd2\x02\x4b\x52\x49\x31\x3c"
"\x43\x69\x3a\x2d\x8e\x70\xad\x7e"
"\xe0\xe5\x46\x09\x80\x89\x13\xb2"
"\x8c\x8b\xd9\x3f\x86\xfb\xb5\x6b",
.ilen = 32,
.result = "\x39\xdf\x7c\x3c\x5a\x29\xb9\x62"
"\x5d\x51\xc2\x16\xd8\xbd\x06\x9f"
"\x9b\x6a\x09\x70\xc1\x51\x83\xc2"
"\x66\x88\x1d\x4f\x9a\xda\xe0\x1e"
"\xc7\x79\x11\x58\xe5\x6b\x20\x40"
"\x7a\xea\x46\x42\x8b\xe4\x6f\xe1",
.rlen = 48,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe0\x8d\x99\x71\x60\xd7\x97\x1a"
"\xbd\x01\x99\xd5\x8a\xdf\x71\x3a"
"\xd3\xdf\x24\x4b\x5e\x3d\x4b\x4e"
"\x30\x7a\xb9\xd8\x53\x0a\x5e\x2b",
.klen = 32,
.iv = "\x03\x1e\x29\x91\xad\x8e\xc1\x53"
"\x0a\xcf\x2d\xbe\x00\x00\x00\x00",
.assoc = "\x19\xb6\x1f\x57\xc4\xf3\xf0\x8b"
"\x78\x2b\x94\x02\x29\x0f\x42\x27"
"\x6b\x75\xcb\x98\x34\x08\x7e\x79"
"\xe4\x3e\x49\x0d\x84\x8b\x22\x87",
.alen = 32,
.input = "\xe1\xd9\xd8\x13\xeb\x3a\x75\x3f"
"\x9d\xbd\x5f\x66\xbe\xdc\xbb\x66"
"\xbf\x17\x99\x62\x4a\x39\x27\x1f"
"\x1d\xdc\x24\xae\x19\x2f\x98\x4c",
.ilen = 32,
.result = "\x19\xb8\x61\x33\x45\x2b\x43\x96"
"\x6f\x51\xd0\x20\x30\x7d\x9b\xc6"
"\x26\x3d\xf8\xc9\x65\x16\xa8\x9f"
"\xf0\x62\x17\x34\xf2\x1e\x8d\x75"
"\x4e\x13\xcc\xc0\xc3\x2a\x54\x2d",
.rlen = 40,
}, {
.key = "\x7c\xc8\x18\x3b\x8d\x99\xe0\x7c"
"\x45\x41\xb8\xbd\x5c\xa7\xc2\x32"
"\x8a\xb8\x02\x59\xa4\xfe\xa9\x2c"
"\x09\x75\x9a\x9b\x3c\x9b\x27\x39",
.klen = 32,
.iv = "\x03\xf9\xd9\x4e\x63\xb5\x3d\x9d"
"\x43\xf6\x1e\x50\0\0\0\0",
.assoc = "\x57\xf5\x6b\x8b\x57\x5c\x3d\x3b"
"\x13\x02\x01\x0c\x83\x4c\x96\x35"
"\x8e\xd6\x39\xcf\x7d\x14\x9b\x94"
"\xb0\x39\x36\xe6\x8f\x57\xe0\x13",
.alen = 32,
.input = "\x3b\x6c\x29\x36\xb6\xef\x07\xa6"
"\x83\x72\x07\x4f\xcf\xfa\x66\x89"
"\x5f\xca\xb1\xba\xd5\x8f\x2c\x27"
"\x30\xdb\x75\x09\x93\xd4\x65\xe4",
.ilen = 32,
.result = "\xb0\x88\x5a\x33\xaa\xe5\xc7\x1d"
"\x85\x23\xc7\xc6\x2f\xf4\x1e\x3d"
"\xcc\x63\x44\x25\x07\x78\x4f\x9e"
"\x96\xb8\x88\xeb\xbc\x48\x1f\x06"
"\x39\xaf\x39\xac\xd8\x4a\x80\x39"
"\x7b\x72\x8a\xf7",
.rlen = 44,
}, {
.key = "\xab\xd0\xe9\x33\x07\x26\xe5\x83"
"\x8c\x76\x95\xd4\xb6\xdc\xf3\x46"
"\xf9\x8f\xad\xe3\x02\x13\x83\x77"
"\x3f\xb0\xf1\xa1\xa1\x22\x0f\x2b",
.klen = 32,
.iv = "\x03\x24\xa7\x8b\x07\xcb\xcc\x0e"
"\xe6\x33\xbf\xf5\x00\x00\x00\x00",
.assoc = "\xd4\xdb\x30\x1d\x03\xfe\xfd\x5f"
"\x87\xd4\x8c\xb6\xb6\xf1\x7a\x5d"
"\xab\x90\x65\x8d\x8e\xca\x4d\x4f"
"\x16\x0c\x40\x90\x4b\xc7\x36\x73",
.alen = 32,
.input = "\xf5\xc6\x7d\x48\xc1\xb7\xe6\x92"
"\x97\x5a\xca\xc4\xa9\x6d\xf9\x3d"
"\x6c\xde\xbc\xf1\x90\xea\x6a\xb2"
"\x35\x86\x36\xaf\x5c\xfe\x4b\x3a",
.ilen = 32,
.result = "\x83\x6f\x40\x87\x72\xcf\xc1\x13"
"\xef\xbb\x80\x21\x04\x6c\x58\x09"
"\x07\x1b\xfc\xdf\xc0\x3f\x5b\xc7"
"\xe0\x79\xa8\x6e\x71\x7c\x3f\xcf"
"\x5c\xda\xb2\x33\xe5\x13\xe2\x0d"
"\x74\xd1\xef\xb5\x0f\x3a\xb5\xf8",
.rlen = 48,
}
};
static const struct aead_testvec aes_ccm_dec_tv_template[] = {
{ /* From RFC 3610 */
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x03\x02\x01\x00"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07",
.alen = 8,
.input = "\x58\x8c\x97\x9a\x61\xc6\x63\xd2"
"\xf0\x66\xd0\xc2\xc0\xf9\x89\x80"
"\x6d\x5f\x6b\x61\xda\xc3\x84\x17"
"\xe8\xd1\x2c\xfd\xf9\x26\xe0",
.ilen = 31,
.result = "\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e",
.rlen = 23,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x07\x06\x05\x04"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b",
.alen = 12,
.input = "\xdc\xf1\xfb\x7b\x5d\x9e\x23\xfb"
"\x9d\x4e\x13\x12\x53\x65\x8a\xd8"
"\x6e\xbd\xca\x3e\x51\xe8\x3f\x07"
"\x7d\x9c\x2d\x93",
.ilen = 28,
.result = "\x0c\x0d\x0e\x0f\x10\x11\x12\x13"
"\x14\x15\x16\x17\x18\x19\x1a\x1b"
"\x1c\x1d\x1e\x1f",
.rlen = 20,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x0b\x0a\x09\x08"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07",
.alen = 8,
.input = "\x82\x53\x1a\x60\xcc\x24\x94\x5a"
"\x4b\x82\x79\x18\x1a\xb5\xc8\x4d"
"\xf2\x1c\xe7\xf9\xb7\x3f\x42\xe1"
"\x97\xea\x9c\x07\xe5\x6b\x5e\xb1"
"\x7e\x5f\x4e",
.ilen = 35,
.result = "\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20",
.rlen = 25,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x00\x00\x0c\x0b\x0a\x09"
"\xa0\xa1\xa2\xa3\xa4\xa5\x00\x00",
.assoc = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b",
.alen = 12,
.input = "\x07\x34\x25\x94\x15\x77\x85\x15"
"\x2b\x07\x40\x98\x33\x0a\xbb\x14"
"\x1b\x94\x7b\x56\x6a\xa9\x40\x6b"
"\x4d\x99\x99\x88\xdd",
.ilen = 29,
.result = "\x0c\x0d\x0e\x0f\x10\x11\x12\x13"
"\x14\x15\x16\x17\x18\x19\x1a\x1b"
"\x1c\x1d\x1e",
.rlen = 19,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\x33\x56\x8e\xf7\xb2\x63"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\x63\x01\x8f\x76\xdc\x8a\x1b\xcb",
.alen = 8,
.input = "\x4c\xcb\x1e\x7c\xa9\x81\xbe\xfa"
"\xa0\x72\x6c\x55\xd3\x78\x06\x12"
"\x98\xc8\x5c\x92\x81\x4a\xbc\x33"
"\xc5\x2e\xe8\x1d\x7d\x77\xc0\x8a",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x90\x20\xea\x6f\x91\xbd\xd8\x5a"
"\xfa\x00\x39\xba\x4b\xaf\xf9\xbf"
"\xb7\x9c\x70\x28\x94\x9c\xd0\xec",
.rlen = 24,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x01\x00\xd5\x60\x91\x2d\x3f\x70"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\xcd\x90\x44\xd2\xb7\x1f\xdb\x81"
"\x20\xea\x60\xc0",
.alen = 12,
.input = "\x00\x97\x69\xec\xab\xdf\x48\x62"
"\x55\x94\xc5\x92\x51\xe6\x03\x57"
"\x22\x67\x5e\x04\xc8\x47\x09\x9e"
"\x5a\xe0\x70\x45\x51",
.ilen = 29,
.result = "\x64\x35\xac\xba\xfb\x11\xa8\x2e"
"\x2f\x07\x1d\x7c\xa4\xa5\xeb\xd9"
"\x3a\x80\x3b\xa8\x7f",
.rlen = 21,
}, {
.key = "\xd7\x82\x8d\x13\xb2\xb0\xbd\xc3"
"\x25\xa7\x62\x36\xdf\x93\xcc\x6b",
.klen = 16,
.iv = "\x01\x00\x42\xff\xf8\xf1\x95\x1c"
"\x3c\x96\x96\x76\x6c\xfa\x00\x00",
.assoc = "\xd8\x5b\xc7\xe6\x9f\x94\x4f\xb8",
.alen = 8,
.input = "\xbc\x21\x8d\xaa\x94\x74\x27\xb6"
"\xdb\x38\x6a\x99\xac\x1a\xef\x23"
"\xad\xe0\xb5\x29\x39\xcb\x6a\x63"
"\x7c\xf9\xbe\xc2\x40\x88\x97\xc6"
"\xba",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 33,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x8a\x19\xb9\x50\xbc\xf7\x1a\x01"
"\x8e\x5e\x67\x01\xc9\x17\x87\x65"
"\x98\x09\xd6\x7d\xbe\xdd\x18",
.rlen = 23,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/* This is taken from FIPS CAVS. */
.key = "\xab\x2f\x8a\x74\xb7\x1c\xd2\xb1"
"\xff\x80\x2e\x48\x7d\x82\xf8\xb9",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x03\xc6\xfb\x7d\x80\x0d\x13\xab"
"\xd8\xa6\xb2\xd8\x00\x00\x00\x00",
.alen = 0,
.input = "\xd5\xe8\x93\x9f\xc7\x89\x2e\x2b",
.ilen = 8,
.result = "\x00",
.rlen = 0,
.novrfy = 1,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xab\x2f\x8a\x74\xb7\x1c\xd2\xb1"
"\xff\x80\x2e\x48\x7d\x82\xf8\xb9",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x03\xaf\x94\x87\x78\x35\x82\x81"
"\x7f\x88\x94\x68\x00\x00\x00\x00",
.alen = 0,
.input = "\x41\x3c\xb8\x87\x73\xcb\xf3\xf3",
.ilen = 8,
.result = "\x00",
.rlen = 0,
}, {
.key = "\x61\x0e\x8c\xae\xe3\x23\xb6\x38"
"\x76\x1c\xf6\x3a\x67\xa3\x9c\xd8",
.klen = 16,
.iv = "\x03\xc6\xfb\x7d\x80\x0d\x13\xab"
"\xd8\xa6\xb2\xd8\x00\x00\x00\x00",
.assoc = "\xf3\x94\x87\x78\x35\x82\x81\x7f"
"\x88\x94\x68\xb1\x78\x6b\x2b\xd6"
"\x04\x1f\x4e\xed\x78\xd5\x33\x66"
"\xd8\x94\x99\x91\x81\x54\x62\x57",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xf0\x7c\x29\x02\xae\x1c\x2f\x55"
"\xd0\xd1\x3d\x1a\xa3\x6d\xe4\x0a"
"\x86\xb0\x87\x6b\x62\x33\x8c\x34"
"\xce\xab\x57\xcc\x79\x0b\xe0\x6f"
"\x5c\x3e\x48\x1f\x6c\x46\xf7\x51"
"\x8b\x84\x83\x2a\xc1\x05\xb8\xc5",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 48,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x50\x82\x3e\x07\xe2\x1e\xb6\xfb"
"\x33\xe4\x73\xce\xd2\xfb\x95\x79"
"\xe8\xb4\xb5\x77\x11\x10\x62\x6f"
"\x6a\x82\xd1\x13\xec\xf5\xd0\x48",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.rlen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.novrfy = 1,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x61\x0e\x8c\xae\xe3\x23\xb6\x38"
"\x76\x1c\xf6\x3a\x67\xa3\x9c\xd8",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x03\x05\xe0\xc9\x0f\xed\x34\xea"
"\x97\xd4\x3b\xdf\x00\x00\x00\x00",
.assoc = "\x49\x5c\x50\x1f\x1d\x94\xcc\x81"
"\xba\xb7\xb6\x03\xaf\xa5\xc1\xa1"
"\xd8\x5c\x42\x68\xe0\x6c\xda\x89"
"\x05\xac\x56\xac\x1b\x2a\xd3\x86",
.alen = 32,
.input = "\x39\xbe\x7d\x15\x62\x77\xf3\x3c"
"\xad\x83\x52\x6d\x71\x03\x25\x1c"
"\xed\x81\x3a\x9a\x16\x7d\x19\x80"
"\x72\x04\x72\xd0\xf6\xff\x05\x0f"
"\xb7\x14\x30\x00\x32\x9e\xa0\xa6"
"\x9e\x5a\x18\xa1\xb8\xfe\xdb\xd3",
.ilen = 48,
.result = "\x75\x05\xbe\xc2\xd9\x1e\xde\x60"
"\x47\x3d\x8c\x7d\xbd\xb5\xd9\xb7"
"\xf2\xae\x61\x05\x8f\x82\x24\x3f"
"\x9c\x67\x91\xe1\x38\x4f\xe4\x0c",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x39\xbb\xa7\xbe\x59\x97\x9e\x73"
"\xa2\xbc\x6b\x98\xd7\x75\x7f\xe3"
"\xa4\x48\x93\x39\x26\x71\x4a\xc6",
.klen = 24,
.iv = "\x03\xee\x49\x83\xe9\xa9\xff\xe9"
"\x57\xba\xfd\x9e\x00\x00\x00\x00",
.assoc = "\x44\xa6\x2c\x05\xe9\xe1\x43\xb1"
"\x58\x7c\xf2\x5c\x6d\x39\x0a\x64"
"\xa4\xf0\x13\x05\xd1\x77\x99\x67"
"\x11\xc4\xc6\xdb\x00\x56\x36\x61",
.alen = 32,
.input = "\x71\x99\xfa\xf4\x44\x12\x68\x9b",
.ilen = 8,
.result = "\x00",
.rlen = 0,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x58\x5d\xa0\x96\x65\x1a\x04\xd7"
"\x96\xe5\xc5\x68\xaa\x95\x35\xe0"
"\x29\xa0\xba\x9e\x48\x78\xd1\xba",
.klen = 24,
.iv = "\x03\xee\x49\x83\xe9\xa9\xff\xe9"
"\x57\xba\xfd\x9e\x00\x00\x00\x00",
.assoc = "\x44\xa6\x2c\x05\xe9\xe1\x43\xb1"
"\x58\x7c\xf2\x5c\x6d\x39\x0a\x64"
"\xa4\xf0\x13\x05\xd1\x77\x99\x67"
"\x11\xc4\xc6\xdb\x00\x56\x36\x61",
.alen = 32,
.input = "\xfb\xe5\x5d\x34\xbe\xe5\xe8\xe7"
"\x5a\xef\x2f\xbf\x1f\x7f\xd4\xb2"
"\x66\xca\x61\x1e\x96\x7a\x61\xb3"
"\x1c\x16\x45\x52\xba\x04\x9c\x9f"
"\xb1\xd2\x40\xbc\x52\x7c\x6f\xb1",
.ilen = 40,
.result = "\x85\x34\x66\x42\xc8\x92\x0f\x36"
"\x58\xe0\x6b\x91\x3c\x98\x5c\xbb"
"\x0a\x85\xcc\x02\xad\x7a\x96\xe9"
"\x65\x43\xa4\xc3\x0f\xdc\x55\x81",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x58\x5d\xa0\x96\x65\x1a\x04\xd7"
"\x96\xe5\xc5\x68\xaa\x95\x35\xe0"
"\x29\xa0\xba\x9e\x48\x78\xd1\xba",
.klen = 24,
.iv = "\x03\xd1\xfc\x57\x9c\xfe\xb8\x9c"
"\xad\x71\xaa\x1f\x00\x00\x00\x00",
.assoc = "\x86\x67\xa5\xa9\x14\x5f\x0d\xc6"
"\xff\x14\xc7\x44\xbf\x6c\x3a\xc3"
"\xff\xb6\x81\xbd\xe2\xd5\x06\xc7"
"\x3c\xa1\x52\x13\x03\x8a\x23\x3a",
.alen = 32,
.input = "\x3f\x66\xb0\x9d\xe5\x4b\x38\x00"
"\xc6\x0e\x6e\xe5\xd6\x98\xa6\x37"
"\x8c\x26\x33\xc6\xb2\xa2\x17\xfa"
"\x64\x19\xc0\x30\xd7\xfc\x14\x6b"
"\xe3\x33\xc2\x04\xb0\x37\xbe\x3f"
"\xa9\xb4\x2d\x68\x03\xa3\x44\xef",
.ilen = 48,
.result = "\x02\x87\x4d\x28\x80\x6e\xb2\xed"
"\x99\x2a\xa8\xca\x04\x25\x45\x90"
"\x1d\xdd\x5a\xd9\xe4\xdb\x9c\x9c"
"\x49\xe9\x01\xfe\xa7\x80\x6d\x6b",
.rlen = 32,
.novrfy = 1,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa4\x4b\x54\x29\x0a\xb8\x6d\x01"
"\x5b\x80\x2a\xcf\x25\xc4\xb7\x5c"
"\x20\x2c\xad\x30\xc2\x2b\x41\xfb"
"\x0e\x85\xbc\x33\xad\x0f\x2b\xff",
.klen = 32,
.iv = "\x03\xee\x49\x83\xe9\xa9\xff\xe9"
"\x57\xba\xfd\x9e\x00\x00\x00\x00",
.alen = 0,
.input = "\x1f\xb8\x8f\xa3\xdd\x54\x00\xf2",
.ilen = 8,
.result = "\x00",
.rlen = 0,
}, {
.key = "\x39\xbb\xa7\xbe\x59\x97\x9e\x73"
"\xa2\xbc\x6b\x98\xd7\x75\x7f\xe3"
"\xa4\x48\x93\x39\x26\x71\x4a\xc6"
"\xae\x8f\x11\x4c\xc2\x9c\x4a\xbb",
.klen = 32,
.iv = "\x03\x85\x34\x66\x42\xc8\x92\x0f"
"\x36\x58\xe0\x6b\x00\x00\x00\x00",
.alen = 0,
.input = "\x48\x01\x5e\x02\x24\x04\x66\x47"
"\xa1\xea\x6f\xaf\xe8\xfc\xfb\xdd"
"\xa5\xa9\x87\x8d\x84\xee\x2e\x77"
"\xbb\x86\xb9\xf5\x5c\x6c\xff\xf6"
"\x72\xc3\x8e\xf7\x70\xb1\xb2\x07"
"\xbc\xa8\xa3\xbd\x83\x7c\x1d\x2a",
.ilen = 48,
.result = "\xdc\x56\xf2\x71\xb0\xb1\xa0\x6c"
"\xf0\x97\x3a\xfb\x6d\xe7\x32\x99"
"\x3e\xaf\x70\x5e\xb2\x4d\xea\x39"
"\x89\xd4\x75\x7a\x63\xb1\xda\x93",
.rlen = 32,
.novrfy = 1,
}, {
.key = "\x58\x5d\xa0\x96\x65\x1a\x04\xd7"
"\x96\xe5\xc5\x68\xaa\x95\x35\xe0"
"\x29\xa0\xba\x9e\x48\x78\xd1\xba"
"\x0d\x1a\x53\x3b\xb5\xe3\xf8\x8b",
.klen = 32,
.iv = "\x03\xcf\x76\x3f\xd9\x95\x75\x8f"
"\x44\x89\x40\x7b\x00\x00\x00\x00",
.assoc = "\x8f\x86\x6c\x4d\x1d\xc5\x39\x88"
"\xc8\xf3\x5c\x52\x10\x63\x6f\x2b"
"\x8a\x2a\xc5\x6f\x30\x23\x58\x7b"
"\xfb\x36\x03\x11\xb4\xd9\xf2\xfe",
.alen = 32,
.input = "\x48\x58\xd6\xf3\xad\x63\x58\xbf"
"\xae\xc7\x5e\xae\x83\x8f\x7b\xe4"
"\x78\x5c\x4c\x67\x71\x89\x94\xbf"
"\x47\xf1\x63\x7e\x1c\x59\xbd\xc5"
"\x7f\x44\x0a\x0c\x01\x18\x07\x92"
"\xe1\xd3\x51\xce\x32\x6d\x0c\x5b",
.ilen = 48,
.result = "\xc2\x54\xc8\xde\x78\x87\x77\x40"
"\x49\x71\xe4\xb7\xe7\xcb\x76\x61"
"\x0a\x41\xb9\xe9\xc0\x76\x54\xab"
"\x04\x49\x3b\x19\x93\x57\x25\x5d",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
},
};
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* rfc4309 refers to section 8 of rfc3610 for test vectors, but they all
* use a 13-byte nonce, we only support an 11-byte nonce. Worse,
* they use AD lengths which are not valid ESP header lengths.
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* These vectors are copied/generated from the ones for rfc4106 with
* the key truncated by one byte..
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aes_ccm_rfc4309_enc_tv_template[] = {
{ /* Generated using Crypto++ */
.key = zeroed_string,
.klen = 19,
.iv = zeroed_string,
.input = zeroed_string,
.ilen = 16,
.assoc = zeroed_string,
.alen = 16,
.result = "\x2E\x9A\xCA\x6B\xDA\x54\xFC\x6F"
"\x12\x50\xE8\xDE\x81\x3C\x63\x08"
"\x1A\x22\xBA\x75\xEE\xD4\xD5\xB5"
"\x27\x50\x01\xAC\x03\x33\x39\xFB",
.rlen = 32,
},{
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = zeroed_string,
.ilen = 16,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xCF\xB9\x99\x17\xC8\x86\x0E\x7F"
"\x7E\x76\xF8\xE6\xF8\xCC\x1F\x17"
"\x6A\xE0\x53\x9F\x4B\x73\x7E\xDA"
"\x08\x09\x4E\xC4\x1E\xAD\xC6\xB0",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = zeroed_string,
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 16,
.assoc = zeroed_string,
.alen = 16,
.result = "\x33\xDE\x73\xBC\xA6\xCE\x4E\xA6"
"\x61\xF4\xF5\x41\x03\x4A\xE3\x86"
"\xA1\xE2\xC2\x42\x2B\x81\x70\x40"
"\xFD\x7F\x76\xD1\x03\x07\xBB\x0C",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = zeroed_string,
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.result = "\x33\xDE\x73\xBC\xA6\xCE\x4E\xA6"
"\x61\xF4\xF5\x41\x03\x4A\xE3\x86"
"\x5B\xC0\x73\xE0\x2B\x73\x68\xC9"
"\x2D\x8C\x58\xC2\x90\x3D\xB0\x3E",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 16,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xCE\xB8\x98\x16\xC9\x87\x0F\x7E"
"\x7F\x77\xF9\xE7\xF9\xCD\x1E\x16"
"\x43\x8E\x76\x57\x3B\xB4\x05\xE8"
"\xA9\x9B\xBF\x25\xE0\x4F\xC0\xED",
.rlen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.input = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.ilen = 64,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.result = "\xCE\xB8\x98\x16\xC9\x87\x0F\x7E"
"\x7F\x77\xF9\xE7\xF9\xCD\x1E\x16"
"\x9C\xA4\x97\x83\x3F\x01\xA5\xF4"
"\x43\x09\xE7\xB8\xE9\xD1\xD7\x02"
"\x9B\xAB\x39\x18\xEB\x94\x34\x36"
"\xE6\xC5\xC8\x9B\x00\x81\x9E\x49"
"\x1D\x78\xE1\x48\xE3\xE9\xEA\x8E"
"\x3A\x2B\x67\x5D\x35\x6A\x0F\xDB"
"\x02\x73\xDD\xE7\x30\x4A\x30\x54"
"\x1A\x9D\x09\xCA\xC8\x1C\x32\x5F",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x45\x67\x89\xab\xcd\xef",
.input = "\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff",
.ilen = 192,
.assoc = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\x00\x00\x45\x67"
"\x89\xab\xcd\xef",
.alen = 20,
.result = "\x64\x17\xDC\x24\x9D\x92\xBA\x5E"
"\x7C\x64\x6D\x33\x46\x77\xAC\xB1"
"\x5C\x9E\xE2\xC7\x27\x11\x3E\x95"
"\x7D\xBE\x28\xC8\xC1\xCA\x5E\x8C"
"\xB4\xE2\xDE\x9F\x53\x59\x26\xDB"
"\x0C\xD4\xE4\x07\x9A\xE6\x3E\x01"
"\x58\x0D\x3E\x3D\xD5\x21\xEB\x04"
"\x06\x9D\x5F\xB9\x02\x49\x1A\x2B"
"\xBA\xF0\x4E\x3B\x85\x50\x5B\x09"
"\xFE\xEC\xFC\x54\xEC\x0C\xE2\x79"
"\x8A\x2F\x5F\xD7\x05\x5D\xF1\x6D"
"\x22\xEB\xD1\x09\x80\x3F\x5A\x70"
"\xB2\xB9\xD3\x63\x99\xC2\x4D\x1B"
"\x36\x12\x00\x89\xAA\x5D\x55\xDA"
"\x1D\x5B\xD8\x3C\x5F\x09\xD2\xE6"
"\x39\x41\x5C\xF0\xBE\x26\x4E\x5F"
"\x2B\x50\x44\x52\xC2\x10\x7D\x38"
"\x82\x64\x83\x0C\xAE\x49\xD0\xE5"
"\x4F\xE5\x66\x4C\x58\x7A\xEE\x43"
"\x3B\x51\xFE\xBA\x24\x8A\xFE\xDC"
"\x19\x6D\x60\x66\x61\xF9\x9A\x3F"
"\x75\xFC\x38\x53\x5B\xB5\xCD\x52"
"\x4F\xE5\xE4\xC9\xFE\x10\xCB\x98"
"\xF0\x06\x5B\x07\xAB\xBB\xF4\x0E"
"\x2D\xC2\xDD\x5D\xDD\x22\x9A\xCC"
"\x39\xAB\x63\xA5\x3D\x9C\x51\x8A",
.rlen = 208,
}, { /* From draft-mcgrew-gcm-test-01 */
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x2E\x44\x3B",
.klen = 19,
.iv = "\x49\x56\xED\x7E\x3B\x24\x4C\xFE",
.input = "\x45\x00\x00\x48\x69\x9A\x00\x00"
"\x80\x11\x4D\xB7\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x9B\xF1\x56"
"\x38\xD3\x01\x00\x00\x01\x00\x00"
"\x00\x00\x00\x00\x04\x5F\x73\x69"
"\x70\x04\x5F\x75\x64\x70\x03\x73"
"\x69\x70\x09\x63\x79\x62\x65\x72"
"\x63\x69\x74\x79\x02\x64\x6B\x00"
"\x00\x21\x00\x01\x01\x02\x02\x01",
.ilen = 72,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x00\x49\x56\xED\x7E"
"\x3B\x24\x4C\xFE",
.alen = 20,
.result = "\x89\xBA\x3E\xEF\xE6\xD6\xCF\xDB"
"\x83\x60\xF5\xBA\x3A\x56\x79\xE6"
"\x7E\x0C\x53\xCF\x9E\x87\xE0\x4E"
"\x1A\x26\x01\x24\xC7\x2E\x3D\xBF"
"\x29\x2C\x91\xC1\xB8\xA8\xCF\xE0"
"\x39\xF8\x53\x6D\x31\x22\x2B\xBF"
"\x98\x81\xFC\x34\xEE\x85\x36\xCD"
"\x26\xDB\x6C\x7A\x0C\x77\x8A\x35"
"\x18\x85\x54\xB2\xBC\xDD\x3F\x43"
"\x61\x06\x8A\xDF\x86\x3F\xB4\xAC"
"\x97\xDC\xBD\xFD\x92\x10\xC5\xFF",
.rlen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xCA\xFE\xBA",
.klen = 19,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.input = "\x45\x00\x00\x3E\x69\x8F\x00\x00"
"\x80\x11\x4D\xCC\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x98\x00\x35"
"\x00\x2A\x23\x43\xB2\xD0\x01\x00"
"\x00\x01\x00\x00\x00\x00\x00\x00"
"\x03\x73\x69\x70\x09\x63\x79\x62"
"\x65\x72\x63\x69\x74\x79\x02\x64"
"\x6B\x00\x00\x01\x00\x01\x00\x01",
.ilen = 64,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.alen = 16,
.result = "\x4B\xC2\x70\x60\x64\xD2\xF3\xC8"
"\xE5\x26\x8A\xDE\xB8\x7E\x7D\x16"
"\x56\xC7\xD2\x88\xBA\x8D\x58\xAF"
"\xF5\x71\xB6\x37\x84\xA7\xB1\x99"
"\x51\x5C\x0D\xA0\x27\xDE\xE7\x2D"
"\xEF\x25\x88\x1F\x1D\x77\x11\xFF"
"\xDB\xED\xEE\x56\x16\xC5\x5C\x9B"
"\x00\x62\x1F\x68\x4E\x7C\xA0\x97"
"\x10\x72\x7E\x53\x13\x3B\x68\xE4"
"\x30\x99\x91\x79\x09\xEA\xFF\x6A",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x11\x22\x33",
.klen = 35,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.input = "\x45\x00\x00\x30\x69\xA6\x40\x00"
"\x80\x06\x26\x90\xC0\xA8\x01\x02"
"\x93\x89\x15\x5E\x0A\x9E\x00\x8B"
"\x2D\xC5\x7E\xE0\x00\x00\x00\x00"
"\x70\x02\x40\x00\x20\xBF\x00\x00"
"\x02\x04\x05\xB4\x01\x01\x04\x02"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x4A\x2C\xBF\xE3\x00\x00\x00\x02"
"\x01\x02\x03\x04\x05\x06\x07\x08",
.alen = 16,
.result = "\xD6\x31\x0D\x2B\x3D\x6F\xBD\x2F"
"\x58\x41\x7E\xFF\x9A\x9E\x09\xB4"
"\x1A\xF7\xF6\x42\x31\xCD\xBF\xAD"
"\x27\x0E\x2C\xF2\xDB\x10\xDF\x55"
"\x8F\x0D\xD7\xAC\x23\xBD\x42\x10"
"\xD0\xB2\xAF\xD8\x37\xAC\x6B\x0B"
"\x11\xD4\x0B\x12\xEC\xB4\xB1\x92"
"\x23\xA6\x10\xB0\x26\xD6\xD9\x26"
"\x5A\x48\x6A\x3E",
.rlen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.input = "\x45\x00\x00\x3C\x99\xC5\x00\x00"
"\x80\x01\xCB\x7A\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x07\x5C"
"\x02\x00\x44\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.ilen = 64,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x6B\x9A\xCA\x57\x43\x91\xFC\x6F"
"\x92\x51\x23\xA4\xC1\x5B\xF0\x10"
"\xF3\x13\xF4\xF8\xA1\x9A\xB4\xDC"
"\x89\xC8\xF8\x42\x62\x95\xB7\xCB"
"\xB8\xF5\x0F\x1B\x2E\x94\xA2\xA7"
"\xBF\xFB\x8A\x92\x13\x63\xD1\x3C"
"\x08\xF5\xE8\xA6\xAA\xF6\x34\xF9"
"\x42\x05\xAF\xB3\xE7\x9A\xFC\xEE"
"\x36\x25\xC1\x10\x12\x1C\xCA\x82"
"\xEA\xE6\x63\x5A\x57\x28\xA9\x9A",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x3C\x99\xC3\x00\x00"
"\x80\x01\xCB\x7C\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x08\x5C"
"\x02\x00\x43\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.ilen = 64,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\x6A\x6B\x45\x2B\x7C\x67\x52\xF6"
"\x10\x60\x40\x62\x6B\x4F\x97\x8E"
"\x0B\xB2\x22\x97\xCB\x21\xE0\x90"
"\xA2\xE7\xD1\x41\x30\xE4\x4B\x1B"
"\x79\x01\x58\x50\x01\x06\xE1\xE0"
"\x2C\x83\x79\xD3\xDE\x46\x97\x1A"
"\x30\xB8\xE5\xDF\xD7\x12\x56\x75"
"\xD0\x95\xB7\xB8\x91\x42\xF7\xFD"
"\x97\x57\xCA\xC1\x20\xD0\x86\xB9"
"\x66\x9D\xB4\x2B\x96\x22\xAC\x67",
.rlen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x1C\x42\xA2\x00\x00"
"\x80\x01\x44\x1F\x40\x67\x93\xB6"
"\xE0\x00\x00\x02\x0A\x00\xF5\xFF"
"\x01\x02\x02\x01",
.ilen = 28,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\x6A\x6B\x45\x0B\xA7\x06\x52\xF6"
"\x10\x60\xCF\x01\x6B\x4F\x97\x20"
"\xEA\xB3\x23\x94\xC9\x21\x1D\x33"
"\xA1\xE5\x90\x40\x05\x37\x45\x70"
"\xB5\xD6\x09\x0A\x23\x73\x33\xF9"
"\x08\xB4\x22\xE4",
.rlen = 44,
}, {
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\xCA\xFE\xBA",
.klen = 27,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.input = "\x45\x00\x00\x28\xA4\xAD\x40\x00"
"\x40\x06\x78\x80\x0A\x01\x03\x8F"
"\x0A\x01\x06\x12\x80\x23\x06\xB8"
"\xCB\x71\x26\x02\xDD\x6B\xB0\x3E"
"\x50\x10\x16\xD0\x75\x68\x00\x01",
.ilen = 40,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.alen = 16,
.result = "\x05\x22\x15\xD1\x52\x56\x85\x04"
"\xA8\x5C\x5D\x6D\x7E\x6E\xF5\xFA"
"\xEA\x16\x37\x50\xF3\xDF\x84\x3B"
"\x2F\x32\x18\x57\x34\x2A\x8C\x23"
"\x67\xDF\x6D\x35\x7B\x54\x0D\xFB"
"\x34\xA5\x9F\x6C\x48\x30\x1E\x22"
"\xFE\xB1\x22\x17\x17\x8A\xB9\x5B",
.rlen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8",
.klen = 19,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.input = "\x45\x00\x00\x49\x33\xBA\x00\x00"
"\x7F\x11\x91\x06\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xDD\x7B\x80\x03\x02\xD5"
"\x00\x00\x4E\x20\x00\x1E\x8C\x18"
"\xD7\x5B\x81\xDC\x91\xBA\xA0\x47"
"\x6B\x91\xB9\x24\xB2\x80\x38\x9D"
"\x92\xC9\x63\xBA\xC0\x46\xEC\x95"
"\x9B\x62\x66\xC0\x47\x22\xB1\x49"
"\x23\x01\x01\x01",
.ilen = 76,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.result = "\x92\xD0\x53\x79\x33\x38\xD5\xF3"
"\x7D\xE4\x7A\x8E\x86\x03\xC9\x90"
"\x96\x35\xAB\x9C\xFB\xE8\xA3\x76"
"\xE9\xE9\xE2\xD1\x2E\x11\x0E\x00"
"\xFA\xCE\xB5\x9E\x02\xA7\x7B\xEA"
"\x71\x9A\x58\xFB\xA5\x8A\xE1\xB7"
"\x9C\x39\x9D\xE3\xB5\x6E\x69\xE6"
"\x63\xC9\xDB\x05\x69\x51\x12\xAD"
"\x3E\x00\x32\x73\x86\xF2\xEE\xF5"
"\x0F\xE8\x81\x7E\x84\xD3\xC0\x0D"
"\x76\xD6\x55\xC6\xB4\xC2\x34\xC7"
"\x12\x25\x0B\xF9",
.rlen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C",
.klen = 35,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.input = "\x45\x08\x00\x28\x73\x2C\x00\x00"
"\x40\x06\xE9\xF9\x0A\x01\x06\x12"
"\x0A\x01\x03\x8F\x06\xB8\x80\x23"
"\xDD\x6B\xAF\xBE\xCB\x71\x26\x02"
"\x50\x10\x1F\x64\x6D\x54\x00\x01",
.ilen = 40,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.result = "\xCC\x74\xB7\xD3\xB0\x38\x50\x42"
"\x2C\x64\x87\x46\x1E\x34\x10\x05"
"\x29\x6B\xBB\x36\xE9\x69\xAD\x92"
"\x82\xA1\x10\x6A\xEB\x0F\xDC\x7D"
"\x08\xBA\xF3\x91\xCA\xAA\x61\xDA"
"\x62\xF4\x14\x61\x5C\x9D\xB5\xA7"
"\xEE\xD7\xB9\x7E\x87\x99\x9B\x7D",
.rlen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x49\x33\x3E\x00\x00"
"\x7F\x11\x91\x82\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xCB\x45\x80\x03\x02\x5B"
"\x00\x00\x01\xE0\x00\x1E\x8C\x18"
"\xD6\x57\x59\xD5\x22\x84\xA0\x35"
"\x2C\x71\x47\x5C\x88\x80\x39\x1C"
"\x76\x4D\x6E\x5E\xE0\x49\x6B\x32"
"\x5A\xE2\x70\xC0\x38\x99\x49\x39"
"\x15\x01\x01\x01",
.ilen = 76,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\x6A\x6B\x45\x5E\xD6\x9A\x52\xF6"
"\xEF\x70\x1A\x9C\xE8\xD3\x19\x86"
"\xC8\x02\xF0\xB0\x03\x09\xD9\x02"
"\xA0\xD2\x59\x04\xD1\x85\x2A\x24"
"\x1C\x67\x3E\xD8\x68\x72\x06\x94"
"\x97\xBA\x4F\x76\x8D\xB0\x44\x5B"
"\x69\xBF\xD5\xE2\x3D\xF1\x0B\x0C"
"\xC0\xBF\xB1\x8F\x70\x09\x9E\xCE"
"\xA5\xF2\x55\x58\x84\xFA\xF9\xB5"
"\x23\xF4\x84\x40\x74\x14\x8A\x6B"
"\xDB\xD7\x67\xED\xA4\x93\xF3\x47"
"\xCC\xF7\x46\x6F",
.rlen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C",
.klen = 35,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.input = "\x63\x69\x73\x63\x6F\x01\x72\x75"
"\x6C\x65\x73\x01\x74\x68\x65\x01"
"\x6E\x65\x74\x77\x65\x01\x64\x65"
"\x66\x69\x6E\x65\x01\x74\x68\x65"
"\x74\x65\x63\x68\x6E\x6F\x6C\x6F"
"\x67\x69\x65\x73\x01\x74\x68\x61"
"\x74\x77\x69\x6C\x6C\x01\x64\x65"
"\x66\x69\x6E\x65\x74\x6F\x6D\x6F"
"\x72\x72\x6F\x77\x01\x02\x02\x01",
.ilen = 72,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.result = "\xEA\x15\xC4\x98\xAC\x15\x22\x37"
"\x00\x07\x1D\xBE\x60\x5D\x73\x16"
"\x4D\x0F\xCC\xCE\x8A\xD0\x49\xD4"
"\x39\xA3\xD1\xB1\x21\x0A\x92\x1A"
"\x2C\xCF\x8F\x9D\xC9\x91\x0D\xB4"
"\x15\xFC\xBC\xA5\xC5\xBF\x54\xE5"
"\x1C\xC7\x32\x41\x07\x7B\x2C\xB6"
"\x5C\x23\x7C\x93\xEA\xEF\x23\x1C"
"\x73\xF4\xE7\x12\x84\x4C\x37\x0A"
"\x4A\x8F\x06\x37\x48\xF9\xF9\x05"
"\x55\x13\x40\xC3\xD5\x55\x3A\x3D",
.rlen = 88,
}, {
.key = "\x7D\x77\x3D\x00\xC1\x44\xC5\x25"
"\xAC\x61\x9D\x18\xC8\x4A\x3F\x47"
"\xD9\x66\x42",
.klen = 19,
.iv = "\x43\x45\x7E\x91\x82\x44\x3B\xC6",
.input = "\x01\x02\x02\x01",
.ilen = 4,
.assoc = "\x33\x54\x67\xAE\xFF\xFF\xFF\xFF"
"\x43\x45\x7E\x91\x82\x44\x3B\xC6",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x4C\x72\x63\x30\x2F\xE6\x56\xDD"
"\xD0\xD8\x60\x9D\x8B\xEF\x85\x90"
"\xF7\x61\x24\x62",
.rlen = 20,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8",
.klen = 19,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.input = "\x74\x6F\x01\x62\x65\x01\x6F\x72"
"\x01\x6E\x6F\x74\x01\x74\x6F\x01"
"\x62\x65\x00\x01",
.ilen = 20,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.result = "\xA3\xBF\x52\x52\x65\x83\xBA\x81"
"\x03\x9B\x84\xFC\x44\x8C\xBB\x81"
"\x36\xE1\x78\xBB\xA5\x49\x3A\xD0"
"\xF0\x6B\x21\xAF\x98\xC0\x34\xDC"
"\x17\x17\x65\xAD",
.rlen = 36,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6C\x65\x67\x61\x6C\x69\x7A\x65"
"\x6D\x61\x72\x69\x6A\x75\x61\x6E"
"\x61\x61\x6E\x64\x64\x6F\x69\x74"
"\x62\x65\x66\x6F\x72\x65\x69\x61"
"\x74\x75\x72",
.klen = 35,
.iv = "\x33\x30\x21\x69\x67\x65\x74\x6D",
.input = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x79\x6B\x69\x63\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\x33\x30\x21\x69"
"\x67\x65\x74\x6D",
.alen = 20,
.result = "\x96\xFD\x86\xF8\xD1\x98\xFF\x10"
"\xAB\x8C\xDA\x8A\x5A\x08\x38\x1A"
"\x48\x59\x80\x18\x1A\x18\x1A\x04"
"\xC9\x0D\xE3\xE7\x0E\xA4\x0B\x75"
"\x92\x9C\x52\x5C\x0B\xFB\xF8\xAF"
"\x16\xC3\x35\xA8\xE7\xCE\x84\x04"
"\xEB\x40\x6B\x7A\x8E\x75\xBB\x42"
"\xE0\x63\x4B\x21\x44\xA2\x2B\x2B"
"\x39\xDB\xC8\xDC",
.rlen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.input = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.ilen = 52,
.assoc = "\x3F\x7E\xF6\x42\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.result = "\x6A\x6B\x45\x27\x3F\x9E\x52\xF6"
"\x10\x60\x54\x25\xEB\x80\x04\x93"
"\xCA\x1B\x23\x97\xCB\x21\x2E\x01"
"\xA2\xE7\x95\x41\x30\xE4\x4B\x1B"
"\x79\x01\x58\x50\x01\x06\xE1\xE0"
"\x2C\x83\x79\xD3\xDE\x46\x97\x1A"
"\x44\xCC\x90\xBF\x00\x94\x94\x92"
"\x20\x17\x0C\x1B\x55\xDE\x7E\x68"
"\xF4\x95\x5D\x4F",
.rlen = 68,
}, {
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x22\x43\x3C",
.klen = 19,
.iv = "\x48\x55\xEC\x7D\x3A\x23\x4B\xFD",
.input = "\x08\x00\xC6\xCD\x02\x00\x07\x00"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6A\x6B\x6C\x6D\x6E\x6F\x70"
"\x71\x72\x73\x74\x01\x02\x02\x01",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x07\x48\x55\xEC\x7D"
"\x3A\x23\x4B\xFD",
.alen = 20,
.result = "\x67\xE9\x28\xB3\x1C\xA4\x6D\x02"
"\xF0\xB5\x37\xB6\x6B\x2F\xF5\x4F"
"\xF8\xA3\x4C\x53\xB8\x12\x09\xBF"
"\x58\x7D\xCF\x29\xA3\x41\x68\x6B"
"\xCE\xE8\x79\x85\x3C\xB0\x3A\x8F"
"\x16\xB0\xA1\x26\xC9\xBC\xBC\xA6",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.rlen = 48,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}
};
static const struct aead_testvec aes_ccm_rfc4309_dec_tv_template[] = {
{ /* Generated using Crypto++ */
.key = zeroed_string,
.klen = 19,
.iv = zeroed_string,
.result = zeroed_string,
.rlen = 16,
.assoc = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2E\x9A\xCA\x6B\xDA\x54\xFC\x6F"
"\x12\x50\xE8\xDE\x81\x3C\x63\x08"
"\x1A\x22\xBA\x75\xEE\xD4\xD5\xB5"
"\x27\x50\x01\xAC\x03\x33\x39\xFB",
.ilen = 32,
},{
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.result = zeroed_string,
.rlen = 16,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.input = "\xCF\xB9\x99\x17\xC8\x86\x0E\x7F"
"\x7E\x76\xF8\xE6\xF8\xCC\x1F\x17"
"\x6A\xE0\x53\x9F\x4B\x73\x7E\xDA"
"\x08\x09\x4E\xC4\x1E\xAD\xC6\xB0",
.ilen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = zeroed_string,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
.assoc = zeroed_string,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x33\xDE\x73\xBC\xA6\xCE\x4E\xA6"
"\x61\xF4\xF5\x41\x03\x4A\xE3\x86"
"\xA1\xE2\xC2\x42\x2B\x81\x70\x40"
"\xFD\x7F\x76\xD1\x03\x07\xBB\x0C",
.ilen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = zeroed_string,
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x33\xDE\x73\xBC\xA6\xCE\x4E\xA6"
"\x61\xF4\xF5\x41\x03\x4A\xE3\x86"
"\x5B\xC0\x73\xE0\x2B\x73\x68\xC9"
"\x2D\x8C\x58\xC2\x90\x3D\xB0\x3E",
.ilen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 16,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.input = "\xCE\xB8\x98\x16\xC9\x87\x0F\x7E"
"\x7F\x77\xF9\xE7\xF9\xCD\x1E\x16"
"\x43\x8E\x76\x57\x3B\xB4\x05\xE8"
"\xA9\x9B\xBF\x25\xE0\x4F\xC0\xED",
.ilen = 32,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfe\xff\xe9\x92\x86\x65\x73\x1c"
"\x6d\x6a\x8f\x94\x67\x30\x83\x08"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x01",
.result = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01"
"\x01\x01\x01\x01\x01\x01\x01\x01",
.rlen = 64,
.assoc = "\x01\x01\x01\x01\x01\x01\x01\x01"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.alen = 16,
.input = "\xCE\xB8\x98\x16\xC9\x87\x0F\x7E"
"\x7F\x77\xF9\xE7\xF9\xCD\x1E\x16"
"\x9C\xA4\x97\x83\x3F\x01\xA5\xF4"
"\x43\x09\xE7\xB8\xE9\xD1\xD7\x02"
"\x9B\xAB\x39\x18\xEB\x94\x34\x36"
"\xE6\xC5\xC8\x9B\x00\x81\x9E\x49"
"\x1D\x78\xE1\x48\xE3\xE9\xEA\x8E"
"\x3A\x2B\x67\x5D\x35\x6A\x0F\xDB"
"\x02\x73\xDD\xE7\x30\x4A\x30\x54"
"\x1A\x9D\x09\xCA\xC8\x1C\x32\x5F",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x45\x67\x89\xab\xcd\xef",
.result = "\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff"
"\xff\xff\xff\xff\xff\xff\xff\xff",
.rlen = 192,
.assoc = "\xaa\xaa\xaa\xaa\xaa\xaa\xaa\xaa"
"\xaa\xaa\xaa\xaa\x00\x00\x45\x67"
"\x89\xab\xcd\xef",
.alen = 20,
.input = "\x64\x17\xDC\x24\x9D\x92\xBA\x5E"
"\x7C\x64\x6D\x33\x46\x77\xAC\xB1"
"\x5C\x9E\xE2\xC7\x27\x11\x3E\x95"
"\x7D\xBE\x28\xC8\xC1\xCA\x5E\x8C"
"\xB4\xE2\xDE\x9F\x53\x59\x26\xDB"
"\x0C\xD4\xE4\x07\x9A\xE6\x3E\x01"
"\x58\x0D\x3E\x3D\xD5\x21\xEB\x04"
"\x06\x9D\x5F\xB9\x02\x49\x1A\x2B"
"\xBA\xF0\x4E\x3B\x85\x50\x5B\x09"
"\xFE\xEC\xFC\x54\xEC\x0C\xE2\x79"
"\x8A\x2F\x5F\xD7\x05\x5D\xF1\x6D"
"\x22\xEB\xD1\x09\x80\x3F\x5A\x70"
"\xB2\xB9\xD3\x63\x99\xC2\x4D\x1B"
"\x36\x12\x00\x89\xAA\x5D\x55\xDA"
"\x1D\x5B\xD8\x3C\x5F\x09\xD2\xE6"
"\x39\x41\x5C\xF0\xBE\x26\x4E\x5F"
"\x2B\x50\x44\x52\xC2\x10\x7D\x38"
"\x82\x64\x83\x0C\xAE\x49\xD0\xE5"
"\x4F\xE5\x66\x4C\x58\x7A\xEE\x43"
"\x3B\x51\xFE\xBA\x24\x8A\xFE\xDC"
"\x19\x6D\x60\x66\x61\xF9\x9A\x3F"
"\x75\xFC\x38\x53\x5B\xB5\xCD\x52"
"\x4F\xE5\xE4\xC9\xFE\x10\xCB\x98"
"\xF0\x06\x5B\x07\xAB\xBB\xF4\x0E"
"\x2D\xC2\xDD\x5D\xDD\x22\x9A\xCC"
"\x39\xAB\x63\xA5\x3D\x9C\x51\x8A",
.ilen = 208,
}, { /* From draft-mcgrew-gcm-test-01 */
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x2E\x44\x3B",
.klen = 19,
.iv = "\x49\x56\xED\x7E\x3B\x24\x4C\xFE",
.result = "\x45\x00\x00\x48\x69\x9A\x00\x00"
"\x80\x11\x4D\xB7\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x9B\xF1\x56"
"\x38\xD3\x01\x00\x00\x01\x00\x00"
"\x00\x00\x00\x00\x04\x5F\x73\x69"
"\x70\x04\x5F\x75\x64\x70\x03\x73"
"\x69\x70\x09\x63\x79\x62\x65\x72"
"\x63\x69\x74\x79\x02\x64\x6B\x00"
"\x00\x21\x00\x01\x01\x02\x02\x01",
.rlen = 72,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x00\x49\x56\xED\x7E"
"\x3B\x24\x4C\xFE",
.alen = 20,
.input = "\x89\xBA\x3E\xEF\xE6\xD6\xCF\xDB"
"\x83\x60\xF5\xBA\x3A\x56\x79\xE6"
"\x7E\x0C\x53\xCF\x9E\x87\xE0\x4E"
"\x1A\x26\x01\x24\xC7\x2E\x3D\xBF"
"\x29\x2C\x91\xC1\xB8\xA8\xCF\xE0"
"\x39\xF8\x53\x6D\x31\x22\x2B\xBF"
"\x98\x81\xFC\x34\xEE\x85\x36\xCD"
"\x26\xDB\x6C\x7A\x0C\x77\x8A\x35"
"\x18\x85\x54\xB2\xBC\xDD\x3F\x43"
"\x61\x06\x8A\xDF\x86\x3F\xB4\xAC"
"\x97\xDC\xBD\xFD\x92\x10\xC5\xFF",
.ilen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xCA\xFE\xBA",
.klen = 19,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.result = "\x45\x00\x00\x3E\x69\x8F\x00\x00"
"\x80\x11\x4D\xCC\xC0\xA8\x01\x02"
"\xC0\xA8\x01\x01\x0A\x98\x00\x35"
"\x00\x2A\x23\x43\xB2\xD0\x01\x00"
"\x00\x01\x00\x00\x00\x00\x00\x00"
"\x03\x73\x69\x70\x09\x63\x79\x62"
"\x65\x72\x63\x69\x74\x79\x02\x64"
"\x6B\x00\x00\x01\x00\x01\x00\x01",
.rlen = 64,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.alen = 16,
.input = "\x4B\xC2\x70\x60\x64\xD2\xF3\xC8"
"\xE5\x26\x8A\xDE\xB8\x7E\x7D\x16"
"\x56\xC7\xD2\x88\xBA\x8D\x58\xAF"
"\xF5\x71\xB6\x37\x84\xA7\xB1\x99"
"\x51\x5C\x0D\xA0\x27\xDE\xE7\x2D"
"\xEF\x25\x88\x1F\x1D\x77\x11\xFF"
"\xDB\xED\xEE\x56\x16\xC5\x5C\x9B"
"\x00\x62\x1F\x68\x4E\x7C\xA0\x97"
"\x10\x72\x7E\x53\x13\x3B\x68\xE4"
"\x30\x99\x91\x79\x09\xEA\xFF\x6A",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x11\x22\x33",
.klen = 35,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.result = "\x45\x00\x00\x30\x69\xA6\x40\x00"
"\x80\x06\x26\x90\xC0\xA8\x01\x02"
"\x93\x89\x15\x5E\x0A\x9E\x00\x8B"
"\x2D\xC5\x7E\xE0\x00\x00\x00\x00"
"\x70\x02\x40\x00\x20\xBF\x00\x00"
"\x02\x04\x05\xB4\x01\x01\x04\x02"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x4A\x2C\xBF\xE3\x00\x00\x00\x02"
"\x01\x02\x03\x04\x05\x06\x07\x08",
.alen = 16,
.input = "\xD6\x31\x0D\x2B\x3D\x6F\xBD\x2F"
"\x58\x41\x7E\xFF\x9A\x9E\x09\xB4"
"\x1A\xF7\xF6\x42\x31\xCD\xBF\xAD"
"\x27\x0E\x2C\xF2\xDB\x10\xDF\x55"
"\x8F\x0D\xD7\xAC\x23\xBD\x42\x10"
"\xD0\xB2\xAF\xD8\x37\xAC\x6B\x0B"
"\x11\xD4\x0B\x12\xEC\xB4\xB1\x92"
"\x23\xA6\x10\xB0\x26\xD6\xD9\x26"
"\x5A\x48\x6A\x3E",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00",
.klen = 19,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.result = "\x45\x00\x00\x3C\x99\xC5\x00\x00"
"\x80\x01\xCB\x7A\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x07\x5C"
"\x02\x00\x44\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.rlen = 64,
.assoc = "\x00\x00\x00\x00\x00\x00\x00\x01"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.alen = 16,
.input = "\x6B\x9A\xCA\x57\x43\x91\xFC\x6F"
"\x92\x51\x23\xA4\xC1\x5B\xF0\x10"
"\xF3\x13\xF4\xF8\xA1\x9A\xB4\xDC"
"\x89\xC8\xF8\x42\x62\x95\xB7\xCB"
"\xB8\xF5\x0F\x1B\x2E\x94\xA2\xA7"
"\xBF\xFB\x8A\x92\x13\x63\xD1\x3C"
"\x08\xF5\xE8\xA6\xAA\xF6\x34\xF9"
"\x42\x05\xAF\xB3\xE7\x9A\xFC\xEE"
"\x36\x25\xC1\x10\x12\x1C\xCA\x82"
"\xEA\xE6\x63\x5A\x57\x28\xA9\x9A",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x3C\x99\xC3\x00\x00"
"\x80\x01\xCB\x7C\x40\x67\x93\x18"
"\x01\x01\x01\x01\x08\x00\x08\x5C"
"\x02\x00\x43\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x75\x76\x77\x61\x62\x63\x64\x65"
"\x66\x67\x68\x69\x01\x02\x02\x01",
.rlen = 64,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\x6A\x6B\x45\x2B\x7C\x67\x52\xF6"
"\x10\x60\x40\x62\x6B\x4F\x97\x8E"
"\x0B\xB2\x22\x97\xCB\x21\xE0\x90"
"\xA2\xE7\xD1\x41\x30\xE4\x4B\x1B"
"\x79\x01\x58\x50\x01\x06\xE1\xE0"
"\x2C\x83\x79\xD3\xDE\x46\x97\x1A"
"\x30\xB8\xE5\xDF\xD7\x12\x56\x75"
"\xD0\x95\xB7\xB8\x91\x42\xF7\xFD"
"\x97\x57\xCA\xC1\x20\xD0\x86\xB9"
"\x66\x9D\xB4\x2B\x96\x22\xAC\x67",
.ilen = 80,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x1C\x42\xA2\x00\x00"
"\x80\x01\x44\x1F\x40\x67\x93\xB6"
"\xE0\x00\x00\x02\x0A\x00\xF5\xFF"
"\x01\x02\x02\x01",
.rlen = 28,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\x6A\x6B\x45\x0B\xA7\x06\x52\xF6"
"\x10\x60\xCF\x01\x6B\x4F\x97\x20"
"\xEA\xB3\x23\x94\xC9\x21\x1D\x33"
"\xA1\xE5\x90\x40\x05\x37\x45\x70"
"\xB5\xD6\x09\x0A\x23\x73\x33\xF9"
"\x08\xB4\x22\xE4",
.ilen = 44,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\x6D\x6A\x8F\x94\x67\x30\x83\x08"
"\xFE\xFF\xE9\x92\x86\x65\x73\x1C"
"\xCA\xFE\xBA",
.klen = 27,
.iv = "\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
.result = "\x45\x00\x00\x28\xA4\xAD\x40\x00"
"\x40\x06\x78\x80\x0A\x01\x03\x8F"
"\x0A\x01\x06\x12\x80\x23\x06\xB8"
"\xCB\x71\x26\x02\xDD\x6B\xB0\x3E"
"\x50\x10\x16\xD0\x75\x68\x00\x01",
.rlen = 40,
.assoc = "\x00\x00\xA5\xF8\x00\x00\x00\x0A"
"\xFA\xCE\xDB\xAD\xDE\xCA\xF8\x88",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x05\x22\x15\xD1\x52\x56\x85\x04"
"\xA8\x5C\x5D\x6D\x7E\x6E\xF5\xFA"
"\xEA\x16\x37\x50\xF3\xDF\x84\x3B"
"\x2F\x32\x18\x57\x34\x2A\x8C\x23"
"\x67\xDF\x6D\x35\x7B\x54\x0D\xFB"
"\x34\xA5\x9F\x6C\x48\x30\x1E\x22"
"\xFE\xB1\x22\x17\x17\x8A\xB9\x5B",
.ilen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8",
.klen = 19,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.result = "\x45\x00\x00\x49\x33\xBA\x00\x00"
"\x7F\x11\x91\x06\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xDD\x7B\x80\x03\x02\xD5"
"\x00\x00\x4E\x20\x00\x1E\x8C\x18"
"\xD7\x5B\x81\xDC\x91\xBA\xA0\x47"
"\x6B\x91\xB9\x24\xB2\x80\x38\x9D"
"\x92\xC9\x63\xBA\xC0\x46\xEC\x95"
"\x9B\x62\x66\xC0\x47\x22\xB1\x49"
"\x23\x01\x01\x01",
.rlen = 76,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.input = "\x92\xD0\x53\x79\x33\x38\xD5\xF3"
"\x7D\xE4\x7A\x8E\x86\x03\xC9\x90"
"\x96\x35\xAB\x9C\xFB\xE8\xA3\x76"
"\xE9\xE9\xE2\xD1\x2E\x11\x0E\x00"
"\xFA\xCE\xB5\x9E\x02\xA7\x7B\xEA"
"\x71\x9A\x58\xFB\xA5\x8A\xE1\xB7"
"\x9C\x39\x9D\xE3\xB5\x6E\x69\xE6"
"\x63\xC9\xDB\x05\x69\x51\x12\xAD"
"\x3E\x00\x32\x73\x86\xF2\xEE\xF5"
"\x0F\xE8\x81\x7E\x84\xD3\xC0\x0D"
"\x76\xD6\x55\xC6\xB4\xC2\x34\xC7"
"\x12\x25\x0B\xF9",
.ilen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C",
.klen = 35,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.result = "\x45\x08\x00\x28\x73\x2C\x00\x00"
"\x40\x06\xE9\xF9\x0A\x01\x06\x12"
"\x0A\x01\x03\x8F\x06\xB8\x80\x23"
"\xDD\x6B\xAF\xBE\xCB\x71\x26\x02"
"\x50\x10\x1F\x64\x6D\x54\x00\x01",
.rlen = 40,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.input = "\xCC\x74\xB7\xD3\xB0\x38\x50\x42"
"\x2C\x64\x87\x46\x1E\x34\x10\x05"
"\x29\x6B\xBB\x36\xE9\x69\xAD\x92"
"\x82\xA1\x10\x6A\xEB\x0F\xDC\x7D"
"\x08\xBA\xF3\x91\xCA\xAA\x61\xDA"
"\x62\xF4\x14\x61\x5C\x9D\xB5\xA7"
"\xEE\xD7\xB9\x7E\x87\x99\x9B\x7D",
.ilen = 56,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x49\x33\x3E\x00\x00"
"\x7F\x11\x91\x82\xC3\xFB\x1D\x10"
"\xC2\xB1\xD3\x26\xC0\x28\x31\xCE"
"\x00\x35\xCB\x45\x80\x03\x02\x5B"
"\x00\x00\x01\xE0\x00\x1E\x8C\x18"
"\xD6\x57\x59\xD5\x22\x84\xA0\x35"
"\x2C\x71\x47\x5C\x88\x80\x39\x1C"
"\x76\x4D\x6E\x5E\xE0\x49\x6B\x32"
"\x5A\xE2\x70\xC0\x38\x99\x49\x39"
"\x15\x01\x01\x01",
.rlen = 76,
.assoc = "\x42\xF6\x7E\x3F\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\x6A\x6B\x45\x5E\xD6\x9A\x52\xF6"
"\xEF\x70\x1A\x9C\xE8\xD3\x19\x86"
"\xC8\x02\xF0\xB0\x03\x09\xD9\x02"
"\xA0\xD2\x59\x04\xD1\x85\x2A\x24"
"\x1C\x67\x3E\xD8\x68\x72\x06\x94"
"\x97\xBA\x4F\x76\x8D\xB0\x44\x5B"
"\x69\xBF\xD5\xE2\x3D\xF1\x0B\x0C"
"\xC0\xBF\xB1\x8F\x70\x09\x9E\xCE"
"\xA5\xF2\x55\x58\x84\xFA\xF9\xB5"
"\x23\xF4\x84\x40\x74\x14\x8A\x6B"
"\xDB\xD7\x67\xED\xA4\x93\xF3\x47"
"\xCC\xF7\x46\x6F",
.ilen = 92,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\x73\x61\x6C",
.klen = 35,
.iv = "\x61\x6E\x64\x01\x69\x76\x65\x63",
.result = "\x63\x69\x73\x63\x6F\x01\x72\x75"
"\x6C\x65\x73\x01\x74\x68\x65\x01"
"\x6E\x65\x74\x77\x65\x01\x64\x65"
"\x66\x69\x6E\x65\x01\x74\x68\x65"
"\x74\x65\x63\x68\x6E\x6F\x6C\x6F"
"\x67\x69\x65\x73\x01\x74\x68\x61"
"\x74\x77\x69\x6C\x6C\x01\x64\x65"
"\x66\x69\x6E\x65\x74\x6F\x6D\x6F"
"\x72\x72\x6F\x77\x01\x02\x02\x01",
.rlen = 72,
.assoc = "\x17\x40\x5E\x67\x15\x6F\x31\x26"
"\xDD\x0D\xB9\x9B\x61\x6E\x64\x01"
"\x69\x76\x65\x63",
.alen = 20,
.input = "\xEA\x15\xC4\x98\xAC\x15\x22\x37"
"\x00\x07\x1D\xBE\x60\x5D\x73\x16"
"\x4D\x0F\xCC\xCE\x8A\xD0\x49\xD4"
"\x39\xA3\xD1\xB1\x21\x0A\x92\x1A"
"\x2C\xCF\x8F\x9D\xC9\x91\x0D\xB4"
"\x15\xFC\xBC\xA5\xC5\xBF\x54\xE5"
"\x1C\xC7\x32\x41\x07\x7B\x2C\xB6"
"\x5C\x23\x7C\x93\xEA\xEF\x23\x1C"
"\x73\xF4\xE7\x12\x84\x4C\x37\x0A"
"\x4A\x8F\x06\x37\x48\xF9\xF9\x05"
"\x55\x13\x40\xC3\xD5\x55\x3A\x3D",
.ilen = 88,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7D\x77\x3D\x00\xC1\x44\xC5\x25"
"\xAC\x61\x9D\x18\xC8\x4A\x3F\x47"
"\xD9\x66\x42",
.klen = 19,
.iv = "\x43\x45\x7E\x91\x82\x44\x3B\xC6",
.result = "\x01\x02\x02\x01",
.rlen = 4,
.assoc = "\x33\x54\x67\xAE\xFF\xFF\xFF\xFF"
"\x43\x45\x7E\x91\x82\x44\x3B\xC6",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4C\x72\x63\x30\x2F\xE6\x56\xDD"
"\xD0\xD8\x60\x9D\x8B\xEF\x85\x90"
"\xF7\x61\x24\x62",
.ilen = 20,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xAB\xBC\xCD\xDE\xF0\x01\x12\x23"
"\x34\x45\x56\x67\x78\x89\x9A\xAB"
"\xDE\xCA\xF8",
.klen = 19,
.iv = "\xCA\xFE\xDE\xBA\xCE\xFA\xCE\x74",
.result = "\x74\x6F\x01\x62\x65\x01\x6F\x72"
"\x01\x6E\x6F\x74\x01\x74\x6F\x01"
"\x62\x65\x00\x01",
.rlen = 20,
.assoc = "\x00\x00\x01\x00\x00\x00\x00\x00"
"\x00\x00\x00\x01\xCA\xFE\xDE\xBA"
"\xCE\xFA\xCE\x74",
.alen = 20,
.input = "\xA3\xBF\x52\x52\x65\x83\xBA\x81"
"\x03\x9B\x84\xFC\x44\x8C\xBB\x81"
"\x36\xE1\x78\xBB\xA5\x49\x3A\xD0"
"\xF0\x6B\x21\xAF\x98\xC0\x34\xDC"
"\x17\x17\x65\xAD",
.ilen = 36,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6C\x65\x67\x61\x6C\x69\x7A\x65"
"\x6D\x61\x72\x69\x6A\x75\x61\x6E"
"\x61\x61\x6E\x64\x64\x6F\x69\x74"
"\x62\x65\x66\x6F\x72\x65\x69\x61"
"\x74\x75\x72",
.klen = 35,
.iv = "\x33\x30\x21\x69\x67\x65\x74\x6D",
.result = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x79\x6B\x69\x63\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\x33\x30\x21\x69"
"\x67\x65\x74\x6D",
.alen = 20,
.input = "\x96\xFD\x86\xF8\xD1\x98\xFF\x10"
"\xAB\x8C\xDA\x8A\x5A\x08\x38\x1A"
"\x48\x59\x80\x18\x1A\x18\x1A\x04"
"\xC9\x0D\xE3\xE7\x0E\xA4\x0B\x75"
"\x92\x9C\x52\x5C\x0B\xFB\xF8\xAF"
"\x16\xC3\x35\xA8\xE7\xCE\x84\x04"
"\xEB\x40\x6B\x7A\x8E\x75\xBB\x42"
"\xE0\x63\x4B\x21\x44\xA2\x2B\x2B"
"\x39\xDB\xC8\xDC",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3D\xE0\x98\x74\xB3\x88\xE6\x49"
"\x19\x88\xD0\xC3\x60\x7E\xAE\x1F"
"\x57\x69\x0E",
.klen = 19,
.iv = "\x4E\x28\x00\x00\xA2\xFC\xA1\xA3",
.result = "\x45\x00\x00\x30\xDA\x3A\x00\x00"
"\x80\x01\xDF\x3B\xC0\xA8\x00\x05"
"\xC0\xA8\x00\x01\x08\x00\xC6\xCD"
"\x02\x00\x07\x00\x61\x62\x63\x64"
"\x65\x66\x67\x68\x69\x6A\x6B\x6C"
"\x6D\x6E\x6F\x70\x71\x72\x73\x74"
"\x01\x02\x02\x01",
.rlen = 52,
.assoc = "\x3F\x7E\xF6\x42\x10\x10\x10\x10"
"\x10\x10\x10\x10\x4E\x28\x00\x00"
"\xA2\xFC\xA1\xA3",
.alen = 20,
.input = "\x6A\x6B\x45\x27\x3F\x9E\x52\xF6"
"\x10\x60\x54\x25\xEB\x80\x04\x93"
"\xCA\x1B\x23\x97\xCB\x21\x2E\x01"
"\xA2\xE7\x95\x41\x30\xE4\x4B\x1B"
"\x79\x01\x58\x50\x01\x06\xE1\xE0"
"\x2C\x83\x79\xD3\xDE\x46\x97\x1A"
"\x44\xCC\x90\xBF\x00\x94\x94\x92"
"\x20\x17\x0C\x1B\x55\xDE\x7E\x68"
"\xF4\x95\x5D\x4F",
.ilen = 68,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4C\x80\xCD\xEF\xBB\x5D\x10\xDA"
"\x90\x6A\xC7\x3C\x36\x13\xA6\x34"
"\x22\x43\x3C",
.klen = 19,
.iv = "\x48\x55\xEC\x7D\x3A\x23\x4B\xFD",
.result = "\x08\x00\xC6\xCD\x02\x00\x07\x00"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6A\x6B\x6C\x6D\x6E\x6F\x70"
"\x71\x72\x73\x74\x01\x02\x02\x01",
.rlen = 32,
.assoc = "\x00\x00\x43\x21\x87\x65\x43\x21"
"\x00\x00\x00\x07\x48\x55\xEC\x7D"
"\x3A\x23\x4B\xFD",
.alen = 20,
.input = "\x67\xE9\x28\xB3\x1C\xA4\x6D\x02"
"\xF0\xB5\x37\xB6\x6B\x2F\xF5\x4F"
"\xF8\xA3\x4C\x53\xB8\x12\x09\xBF"
"\x58\x7D\xCF\x29\xA3\x41\x68\x6B"
"\xCE\xE8\x79\x85\x3C\xB0\x3A\x8F"
"\x16\xB0\xA1\x26\xC9\xBC\xBC\xA6",
.ilen = 48,
}
};
/*
* ChaCha20-Poly1305 AEAD test vectors from RFC7539 2.8.2./A.5.
*/
static const struct aead_testvec rfc7539_enc_tv_template[] = {
{
.key = "\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x07\x00\x00\x00\x40\x41\x42\x43"
"\x44\x45\x46\x47",
.assoc = "\x50\x51\x52\x53\xc0\xc1\xc2\xc3"
"\xc4\xc5\xc6\xc7",
.alen = 12,
.input = "\x4c\x61\x64\x69\x65\x73\x20\x61"
"\x6e\x64\x20\x47\x65\x6e\x74\x6c"
"\x65\x6d\x65\x6e\x20\x6f\x66\x20"
"\x74\x68\x65\x20\x63\x6c\x61\x73"
"\x73\x20\x6f\x66\x20\x27\x39\x39"
"\x3a\x20\x49\x66\x20\x49\x20\x63"
"\x6f\x75\x6c\x64\x20\x6f\x66\x66"
"\x65\x72\x20\x79\x6f\x75\x20\x6f"
"\x6e\x6c\x79\x20\x6f\x6e\x65\x20"
"\x74\x69\x70\x20\x66\x6f\x72\x20"
"\x74\x68\x65\x20\x66\x75\x74\x75"
"\x72\x65\x2c\x20\x73\x75\x6e\x73"
"\x63\x72\x65\x65\x6e\x20\x77\x6f"
"\x75\x6c\x64\x20\x62\x65\x20\x69"
"\x74\x2e",
.ilen = 114,
.result = "\xd3\x1a\x8d\x34\x64\x8e\x60\xdb"
"\x7b\x86\xaf\xbc\x53\xef\x7e\xc2"
"\xa4\xad\xed\x51\x29\x6e\x08\xfe"
"\xa9\xe2\xb5\xa7\x36\xee\x62\xd6"
"\x3d\xbe\xa4\x5e\x8c\xa9\x67\x12"
"\x82\xfa\xfb\x69\xda\x92\x72\x8b"
"\x1a\x71\xde\x0a\x9e\x06\x0b\x29"
"\x05\xd6\xa5\xb6\x7e\xcd\x3b\x36"
"\x92\xdd\xbd\x7f\x2d\x77\x8b\x8c"
"\x98\x03\xae\xe3\x28\x09\x1b\x58"
"\xfa\xb3\x24\xe4\xfa\xd6\x75\x94"
"\x55\x85\x80\x8b\x48\x31\xd7\xbc"
"\x3f\xf4\xde\xf0\x8e\x4b\x7a\x9d"
"\xe5\x76\xd2\x65\x86\xce\xc6\x4b"
"\x61\x16\x1a\xe1\x0b\x59\x4f\x09"
"\xe2\x6a\x7e\x90\x2e\xcb\xd0\x60"
"\x06\x91",
.rlen = 130,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x00\x00\x00\x01\x02\x03\x04"
"\x05\x06\x07\x08",
.assoc = "\xf3\x33\x88\x86\x00\x00\x00\x00"
"\x00\x00\x4e\x91",
.alen = 12,
.input = "\x49\x6e\x74\x65\x72\x6e\x65\x74"
"\x2d\x44\x72\x61\x66\x74\x73\x20"
"\x61\x72\x65\x20\x64\x72\x61\x66"
"\x74\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x76\x61\x6c\x69"
"\x64\x20\x66\x6f\x72\x20\x61\x20"
"\x6d\x61\x78\x69\x6d\x75\x6d\x20"
"\x6f\x66\x20\x73\x69\x78\x20\x6d"
"\x6f\x6e\x74\x68\x73\x20\x61\x6e"
"\x64\x20\x6d\x61\x79\x20\x62\x65"
"\x20\x75\x70\x64\x61\x74\x65\x64"
"\x2c\x20\x72\x65\x70\x6c\x61\x63"
"\x65\x64\x2c\x20\x6f\x72\x20\x6f"
"\x62\x73\x6f\x6c\x65\x74\x65\x64"
"\x20\x62\x79\x20\x6f\x74\x68\x65"
"\x72\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x61\x74\x20\x61"
"\x6e\x79\x20\x74\x69\x6d\x65\x2e"
"\x20\x49\x74\x20\x69\x73\x20\x69"
"\x6e\x61\x70\x70\x72\x6f\x70\x72"
"\x69\x61\x74\x65\x20\x74\x6f\x20"
"\x75\x73\x65\x20\x49\x6e\x74\x65"
"\x72\x6e\x65\x74\x2d\x44\x72\x61"
"\x66\x74\x73\x20\x61\x73\x20\x72"
"\x65\x66\x65\x72\x65\x6e\x63\x65"
"\x20\x6d\x61\x74\x65\x72\x69\x61"
"\x6c\x20\x6f\x72\x20\x74\x6f\x20"
"\x63\x69\x74\x65\x20\x74\x68\x65"
"\x6d\x20\x6f\x74\x68\x65\x72\x20"
"\x74\x68\x61\x6e\x20\x61\x73\x20"
"\x2f\xe2\x80\x9c\x77\x6f\x72\x6b"
"\x20\x69\x6e\x20\x70\x72\x6f\x67"
"\x72\x65\x73\x73\x2e\x2f\xe2\x80"
"\x9d",
.ilen = 265,
.result = "\x64\xa0\x86\x15\x75\x86\x1a\xf4"
"\x60\xf0\x62\xc7\x9b\xe6\x43\xbd"
"\x5e\x80\x5c\xfd\x34\x5c\xf3\x89"
"\xf1\x08\x67\x0a\xc7\x6c\x8c\xb2"
"\x4c\x6c\xfc\x18\x75\x5d\x43\xee"
"\xa0\x9e\xe9\x4e\x38\x2d\x26\xb0"
"\xbd\xb7\xb7\x3c\x32\x1b\x01\x00"
"\xd4\xf0\x3b\x7f\x35\x58\x94\xcf"
"\x33\x2f\x83\x0e\x71\x0b\x97\xce"
"\x98\xc8\xa8\x4a\xbd\x0b\x94\x81"
"\x14\xad\x17\x6e\x00\x8d\x33\xbd"
"\x60\xf9\x82\xb1\xff\x37\xc8\x55"
"\x97\x97\xa0\x6e\xf4\xf0\xef\x61"
"\xc1\x86\x32\x4e\x2b\x35\x06\x38"
"\x36\x06\x90\x7b\x6a\x7c\x02\xb0"
"\xf9\xf6\x15\x7b\x53\xc8\x67\xe4"
"\xb9\x16\x6c\x76\x7b\x80\x4d\x46"
"\xa5\x9b\x52\x16\xcd\xe7\xa4\xe9"
"\x90\x40\xc5\xa4\x04\x33\x22\x5e"
"\xe2\x82\xa1\xb0\xa0\x6c\x52\x3e"
"\xaf\x45\x34\xd7\xf8\x3f\xa1\x15"
"\x5b\x00\x47\x71\x8c\xbc\x54\x6a"
"\x0d\x07\x2b\x04\xb3\x56\x4e\xea"
"\x1b\x42\x22\x73\xf5\x48\x27\x1a"
"\x0b\xb2\x31\x60\x53\xfa\x76\x99"
"\x19\x55\xeb\xd6\x31\x59\x43\x4e"
"\xce\xbb\x4e\x46\x6d\xae\x5a\x10"
"\x73\xa6\x72\x76\x27\x09\x7a\x10"
"\x49\xe6\x17\xd9\x1d\x36\x10\x94"
"\xfa\x68\xf0\xff\x77\x98\x71\x30"
"\x30\x5b\xea\xba\x2e\xda\x04\xdf"
"\x99\x7b\x71\x4d\x6c\x6f\x2c\x29"
"\xa6\xad\x5c\xb4\x02\x2b\x02\x70"
"\x9b\xee\xad\x9d\x67\x89\x0c\xbb"
"\x22\x39\x23\x36\xfe\xa1\x85\x1f"
"\x38",
.rlen = 281,
},
};
static const struct aead_testvec rfc7539_dec_tv_template[] = {
{
.key = "\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f",
.klen = 32,
.iv = "\x07\x00\x00\x00\x40\x41\x42\x43"
"\x44\x45\x46\x47",
.assoc = "\x50\x51\x52\x53\xc0\xc1\xc2\xc3"
"\xc4\xc5\xc6\xc7",
.alen = 12,
.input = "\xd3\x1a\x8d\x34\x64\x8e\x60\xdb"
"\x7b\x86\xaf\xbc\x53\xef\x7e\xc2"
"\xa4\xad\xed\x51\x29\x6e\x08\xfe"
"\xa9\xe2\xb5\xa7\x36\xee\x62\xd6"
"\x3d\xbe\xa4\x5e\x8c\xa9\x67\x12"
"\x82\xfa\xfb\x69\xda\x92\x72\x8b"
"\x1a\x71\xde\x0a\x9e\x06\x0b\x29"
"\x05\xd6\xa5\xb6\x7e\xcd\x3b\x36"
"\x92\xdd\xbd\x7f\x2d\x77\x8b\x8c"
"\x98\x03\xae\xe3\x28\x09\x1b\x58"
"\xfa\xb3\x24\xe4\xfa\xd6\x75\x94"
"\x55\x85\x80\x8b\x48\x31\xd7\xbc"
"\x3f\xf4\xde\xf0\x8e\x4b\x7a\x9d"
"\xe5\x76\xd2\x65\x86\xce\xc6\x4b"
"\x61\x16\x1a\xe1\x0b\x59\x4f\x09"
"\xe2\x6a\x7e\x90\x2e\xcb\xd0\x60"
"\x06\x91",
.ilen = 130,
.result = "\x4c\x61\x64\x69\x65\x73\x20\x61"
"\x6e\x64\x20\x47\x65\x6e\x74\x6c"
"\x65\x6d\x65\x6e\x20\x6f\x66\x20"
"\x74\x68\x65\x20\x63\x6c\x61\x73"
"\x73\x20\x6f\x66\x20\x27\x39\x39"
"\x3a\x20\x49\x66\x20\x49\x20\x63"
"\x6f\x75\x6c\x64\x20\x6f\x66\x66"
"\x65\x72\x20\x79\x6f\x75\x20\x6f"
"\x6e\x6c\x79\x20\x6f\x6e\x65\x20"
"\x74\x69\x70\x20\x66\x6f\x72\x20"
"\x74\x68\x65\x20\x66\x75\x74\x75"
"\x72\x65\x2c\x20\x73\x75\x6e\x73"
"\x63\x72\x65\x65\x6e\x20\x77\x6f"
"\x75\x6c\x64\x20\x62\x65\x20\x69"
"\x74\x2e",
.rlen = 114,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0",
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x00\x00\x00\x00\x01\x02\x03\x04"
"\x05\x06\x07\x08",
.assoc = "\xf3\x33\x88\x86\x00\x00\x00\x00"
"\x00\x00\x4e\x91",
.alen = 12,
.input = "\x64\xa0\x86\x15\x75\x86\x1a\xf4"
"\x60\xf0\x62\xc7\x9b\xe6\x43\xbd"
"\x5e\x80\x5c\xfd\x34\x5c\xf3\x89"
"\xf1\x08\x67\x0a\xc7\x6c\x8c\xb2"
"\x4c\x6c\xfc\x18\x75\x5d\x43\xee"
"\xa0\x9e\xe9\x4e\x38\x2d\x26\xb0"
"\xbd\xb7\xb7\x3c\x32\x1b\x01\x00"
"\xd4\xf0\x3b\x7f\x35\x58\x94\xcf"
"\x33\x2f\x83\x0e\x71\x0b\x97\xce"
"\x98\xc8\xa8\x4a\xbd\x0b\x94\x81"
"\x14\xad\x17\x6e\x00\x8d\x33\xbd"
"\x60\xf9\x82\xb1\xff\x37\xc8\x55"
"\x97\x97\xa0\x6e\xf4\xf0\xef\x61"
"\xc1\x86\x32\x4e\x2b\x35\x06\x38"
"\x36\x06\x90\x7b\x6a\x7c\x02\xb0"
"\xf9\xf6\x15\x7b\x53\xc8\x67\xe4"
"\xb9\x16\x6c\x76\x7b\x80\x4d\x46"
"\xa5\x9b\x52\x16\xcd\xe7\xa4\xe9"
"\x90\x40\xc5\xa4\x04\x33\x22\x5e"
"\xe2\x82\xa1\xb0\xa0\x6c\x52\x3e"
"\xaf\x45\x34\xd7\xf8\x3f\xa1\x15"
"\x5b\x00\x47\x71\x8c\xbc\x54\x6a"
"\x0d\x07\x2b\x04\xb3\x56\x4e\xea"
"\x1b\x42\x22\x73\xf5\x48\x27\x1a"
"\x0b\xb2\x31\x60\x53\xfa\x76\x99"
"\x19\x55\xeb\xd6\x31\x59\x43\x4e"
"\xce\xbb\x4e\x46\x6d\xae\x5a\x10"
"\x73\xa6\x72\x76\x27\x09\x7a\x10"
"\x49\xe6\x17\xd9\x1d\x36\x10\x94"
"\xfa\x68\xf0\xff\x77\x98\x71\x30"
"\x30\x5b\xea\xba\x2e\xda\x04\xdf"
"\x99\x7b\x71\x4d\x6c\x6f\x2c\x29"
"\xa6\xad\x5c\xb4\x02\x2b\x02\x70"
"\x9b\xee\xad\x9d\x67\x89\x0c\xbb"
"\x22\x39\x23\x36\xfe\xa1\x85\x1f"
"\x38",
.ilen = 281,
.result = "\x49\x6e\x74\x65\x72\x6e\x65\x74"
"\x2d\x44\x72\x61\x66\x74\x73\x20"
"\x61\x72\x65\x20\x64\x72\x61\x66"
"\x74\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x76\x61\x6c\x69"
"\x64\x20\x66\x6f\x72\x20\x61\x20"
"\x6d\x61\x78\x69\x6d\x75\x6d\x20"
"\x6f\x66\x20\x73\x69\x78\x20\x6d"
"\x6f\x6e\x74\x68\x73\x20\x61\x6e"
"\x64\x20\x6d\x61\x79\x20\x62\x65"
"\x20\x75\x70\x64\x61\x74\x65\x64"
"\x2c\x20\x72\x65\x70\x6c\x61\x63"
"\x65\x64\x2c\x20\x6f\x72\x20\x6f"
"\x62\x73\x6f\x6c\x65\x74\x65\x64"
"\x20\x62\x79\x20\x6f\x74\x68\x65"
"\x72\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x61\x74\x20\x61"
"\x6e\x79\x20\x74\x69\x6d\x65\x2e"
"\x20\x49\x74\x20\x69\x73\x20\x69"
"\x6e\x61\x70\x70\x72\x6f\x70\x72"
"\x69\x61\x74\x65\x20\x74\x6f\x20"
"\x75\x73\x65\x20\x49\x6e\x74\x65"
"\x72\x6e\x65\x74\x2d\x44\x72\x61"
"\x66\x74\x73\x20\x61\x73\x20\x72"
"\x65\x66\x65\x72\x65\x6e\x63\x65"
"\x20\x6d\x61\x74\x65\x72\x69\x61"
"\x6c\x20\x6f\x72\x20\x74\x6f\x20"
"\x63\x69\x74\x65\x20\x74\x68\x65"
"\x6d\x20\x6f\x74\x68\x65\x72\x20"
"\x74\x68\x61\x6e\x20\x61\x73\x20"
"\x2f\xe2\x80\x9c\x77\x6f\x72\x6b"
"\x20\x69\x6e\x20\x70\x72\x6f\x67"
"\x72\x65\x73\x73\x2e\x2f\xe2\x80"
"\x9d",
.rlen = 265,
},
};
/*
* draft-irtf-cfrg-chacha20-poly1305
*/
static const struct aead_testvec rfc7539esp_enc_tv_template[] = {
{
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0"
"\x00\x00\x00\x00",
.klen = 36,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.assoc = "\xf3\x33\x88\x86\x00\x00\x00\x00"
"\x00\x00\x4e\x91\x01\x02\x03\x04"
"\x05\x06\x07\x08",
.alen = 20,
.input = "\x49\x6e\x74\x65\x72\x6e\x65\x74"
"\x2d\x44\x72\x61\x66\x74\x73\x20"
"\x61\x72\x65\x20\x64\x72\x61\x66"
"\x74\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x76\x61\x6c\x69"
"\x64\x20\x66\x6f\x72\x20\x61\x20"
"\x6d\x61\x78\x69\x6d\x75\x6d\x20"
"\x6f\x66\x20\x73\x69\x78\x20\x6d"
"\x6f\x6e\x74\x68\x73\x20\x61\x6e"
"\x64\x20\x6d\x61\x79\x20\x62\x65"
"\x20\x75\x70\x64\x61\x74\x65\x64"
"\x2c\x20\x72\x65\x70\x6c\x61\x63"
"\x65\x64\x2c\x20\x6f\x72\x20\x6f"
"\x62\x73\x6f\x6c\x65\x74\x65\x64"
"\x20\x62\x79\x20\x6f\x74\x68\x65"
"\x72\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x61\x74\x20\x61"
"\x6e\x79\x20\x74\x69\x6d\x65\x2e"
"\x20\x49\x74\x20\x69\x73\x20\x69"
"\x6e\x61\x70\x70\x72\x6f\x70\x72"
"\x69\x61\x74\x65\x20\x74\x6f\x20"
"\x75\x73\x65\x20\x49\x6e\x74\x65"
"\x72\x6e\x65\x74\x2d\x44\x72\x61"
"\x66\x74\x73\x20\x61\x73\x20\x72"
"\x65\x66\x65\x72\x65\x6e\x63\x65"
"\x20\x6d\x61\x74\x65\x72\x69\x61"
"\x6c\x20\x6f\x72\x20\x74\x6f\x20"
"\x63\x69\x74\x65\x20\x74\x68\x65"
"\x6d\x20\x6f\x74\x68\x65\x72\x20"
"\x74\x68\x61\x6e\x20\x61\x73\x20"
"\x2f\xe2\x80\x9c\x77\x6f\x72\x6b"
"\x20\x69\x6e\x20\x70\x72\x6f\x67"
"\x72\x65\x73\x73\x2e\x2f\xe2\x80"
"\x9d",
.ilen = 265,
.result = "\x64\xa0\x86\x15\x75\x86\x1a\xf4"
"\x60\xf0\x62\xc7\x9b\xe6\x43\xbd"
"\x5e\x80\x5c\xfd\x34\x5c\xf3\x89"
"\xf1\x08\x67\x0a\xc7\x6c\x8c\xb2"
"\x4c\x6c\xfc\x18\x75\x5d\x43\xee"
"\xa0\x9e\xe9\x4e\x38\x2d\x26\xb0"
"\xbd\xb7\xb7\x3c\x32\x1b\x01\x00"
"\xd4\xf0\x3b\x7f\x35\x58\x94\xcf"
"\x33\x2f\x83\x0e\x71\x0b\x97\xce"
"\x98\xc8\xa8\x4a\xbd\x0b\x94\x81"
"\x14\xad\x17\x6e\x00\x8d\x33\xbd"
"\x60\xf9\x82\xb1\xff\x37\xc8\x55"
"\x97\x97\xa0\x6e\xf4\xf0\xef\x61"
"\xc1\x86\x32\x4e\x2b\x35\x06\x38"
"\x36\x06\x90\x7b\x6a\x7c\x02\xb0"
"\xf9\xf6\x15\x7b\x53\xc8\x67\xe4"
"\xb9\x16\x6c\x76\x7b\x80\x4d\x46"
"\xa5\x9b\x52\x16\xcd\xe7\xa4\xe9"
"\x90\x40\xc5\xa4\x04\x33\x22\x5e"
"\xe2\x82\xa1\xb0\xa0\x6c\x52\x3e"
"\xaf\x45\x34\xd7\xf8\x3f\xa1\x15"
"\x5b\x00\x47\x71\x8c\xbc\x54\x6a"
"\x0d\x07\x2b\x04\xb3\x56\x4e\xea"
"\x1b\x42\x22\x73\xf5\x48\x27\x1a"
"\x0b\xb2\x31\x60\x53\xfa\x76\x99"
"\x19\x55\xeb\xd6\x31\x59\x43\x4e"
"\xce\xbb\x4e\x46\x6d\xae\x5a\x10"
"\x73\xa6\x72\x76\x27\x09\x7a\x10"
"\x49\xe6\x17\xd9\x1d\x36\x10\x94"
"\xfa\x68\xf0\xff\x77\x98\x71\x30"
"\x30\x5b\xea\xba\x2e\xda\x04\xdf"
"\x99\x7b\x71\x4d\x6c\x6f\x2c\x29"
"\xa6\xad\x5c\xb4\x02\x2b\x02\x70"
"\x9b\xee\xad\x9d\x67\x89\x0c\xbb"
"\x22\x39\x23\x36\xfe\xa1\x85\x1f"
"\x38",
.rlen = 281,
},
};
static const struct aead_testvec rfc7539esp_dec_tv_template[] = {
{
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0"
"\x00\x00\x00\x00",
.klen = 36,
.iv = "\x01\x02\x03\x04\x05\x06\x07\x08",
.assoc = "\xf3\x33\x88\x86\x00\x00\x00\x00"
"\x00\x00\x4e\x91\x01\x02\x03\x04"
"\x05\x06\x07\x08",
.alen = 20,
.input = "\x64\xa0\x86\x15\x75\x86\x1a\xf4"
"\x60\xf0\x62\xc7\x9b\xe6\x43\xbd"
"\x5e\x80\x5c\xfd\x34\x5c\xf3\x89"
"\xf1\x08\x67\x0a\xc7\x6c\x8c\xb2"
"\x4c\x6c\xfc\x18\x75\x5d\x43\xee"
"\xa0\x9e\xe9\x4e\x38\x2d\x26\xb0"
"\xbd\xb7\xb7\x3c\x32\x1b\x01\x00"
"\xd4\xf0\x3b\x7f\x35\x58\x94\xcf"
"\x33\x2f\x83\x0e\x71\x0b\x97\xce"
"\x98\xc8\xa8\x4a\xbd\x0b\x94\x81"
"\x14\xad\x17\x6e\x00\x8d\x33\xbd"
"\x60\xf9\x82\xb1\xff\x37\xc8\x55"
"\x97\x97\xa0\x6e\xf4\xf0\xef\x61"
"\xc1\x86\x32\x4e\x2b\x35\x06\x38"
"\x36\x06\x90\x7b\x6a\x7c\x02\xb0"
"\xf9\xf6\x15\x7b\x53\xc8\x67\xe4"
"\xb9\x16\x6c\x76\x7b\x80\x4d\x46"
"\xa5\x9b\x52\x16\xcd\xe7\xa4\xe9"
"\x90\x40\xc5\xa4\x04\x33\x22\x5e"
"\xe2\x82\xa1\xb0\xa0\x6c\x52\x3e"
"\xaf\x45\x34\xd7\xf8\x3f\xa1\x15"
"\x5b\x00\x47\x71\x8c\xbc\x54\x6a"
"\x0d\x07\x2b\x04\xb3\x56\x4e\xea"
"\x1b\x42\x22\x73\xf5\x48\x27\x1a"
"\x0b\xb2\x31\x60\x53\xfa\x76\x99"
"\x19\x55\xeb\xd6\x31\x59\x43\x4e"
"\xce\xbb\x4e\x46\x6d\xae\x5a\x10"
"\x73\xa6\x72\x76\x27\x09\x7a\x10"
"\x49\xe6\x17\xd9\x1d\x36\x10\x94"
"\xfa\x68\xf0\xff\x77\x98\x71\x30"
"\x30\x5b\xea\xba\x2e\xda\x04\xdf"
"\x99\x7b\x71\x4d\x6c\x6f\x2c\x29"
"\xa6\xad\x5c\xb4\x02\x2b\x02\x70"
"\x9b\xee\xad\x9d\x67\x89\x0c\xbb"
"\x22\x39\x23\x36\xfe\xa1\x85\x1f"
"\x38",
.ilen = 281,
.result = "\x49\x6e\x74\x65\x72\x6e\x65\x74"
"\x2d\x44\x72\x61\x66\x74\x73\x20"
"\x61\x72\x65\x20\x64\x72\x61\x66"
"\x74\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x76\x61\x6c\x69"
"\x64\x20\x66\x6f\x72\x20\x61\x20"
"\x6d\x61\x78\x69\x6d\x75\x6d\x20"
"\x6f\x66\x20\x73\x69\x78\x20\x6d"
"\x6f\x6e\x74\x68\x73\x20\x61\x6e"
"\x64\x20\x6d\x61\x79\x20\x62\x65"
"\x20\x75\x70\x64\x61\x74\x65\x64"
"\x2c\x20\x72\x65\x70\x6c\x61\x63"
"\x65\x64\x2c\x20\x6f\x72\x20\x6f"
"\x62\x73\x6f\x6c\x65\x74\x65\x64"
"\x20\x62\x79\x20\x6f\x74\x68\x65"
"\x72\x20\x64\x6f\x63\x75\x6d\x65"
"\x6e\x74\x73\x20\x61\x74\x20\x61"
"\x6e\x79\x20\x74\x69\x6d\x65\x2e"
"\x20\x49\x74\x20\x69\x73\x20\x69"
"\x6e\x61\x70\x70\x72\x6f\x70\x72"
"\x69\x61\x74\x65\x20\x74\x6f\x20"
"\x75\x73\x65\x20\x49\x6e\x74\x65"
"\x72\x6e\x65\x74\x2d\x44\x72\x61"
"\x66\x74\x73\x20\x61\x73\x20\x72"
"\x65\x66\x65\x72\x65\x6e\x63\x65"
"\x20\x6d\x61\x74\x65\x72\x69\x61"
"\x6c\x20\x6f\x72\x20\x74\x6f\x20"
"\x63\x69\x74\x65\x20\x74\x68\x65"
"\x6d\x20\x6f\x74\x68\x65\x72\x20"
"\x74\x68\x61\x6e\x20\x61\x73\x20"
"\x2f\xe2\x80\x9c\x77\x6f\x72\x6b"
"\x20\x69\x6e\x20\x70\x72\x6f\x67"
"\x72\x65\x73\x73\x2e\x2f\xe2\x80"
"\x9d",
.rlen = 265,
crypto: testmgr - Add test vectors for AEGIS This patch adds test vectors for the AEGIS family of AEAD algorithms (AEGIS-128, AEGIS-128L, and AEGIS-256). The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:12:50 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aegis128_enc_tv_template[] = {
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x07\xa5\x11\xf2\x9d\x40\xb8\x6d"
"\xda\xb8\x12\x34\x4c\x53\xd9\x72",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x79",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x9e\x78\x52\xae\xcb\x9e\xe4\xd3"
"\x9a\xd7\x5d\xd7\xaa\x9a\xe9\x5a"
"\xcc",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 15,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xc3\x80\x83\x04\x5f\xaa\x61\xc7"
"\xca\xdd\x6f\xac\x85\x08\xb5\x35"
"\x2b\xc2\x3e\x0b\x1b\x39\x37\x2b"
"\x7a\x21\x16\xb3\xe6\x67\x66",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x23\x25\x30\xe5\x6a\xb6\x36\x7d"
"\x38\xfd\x3a\xd2\xc2\x58\xa9\x11"
"\x1e\xa8\x30\x9c\x16\xa4\xdb\x65"
"\x51\x10\x16\x27\x70\x9b\x64\x29",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x2a\x8d\x56\x91\xc6\xf3\x56\xa5"
"\x1f\xf0\x89\x2e\x13\xad\xe6\xf6"
"\x46\x80\xb1\x0e\x18\x30\x40\x97"
"\x03\xdf\x64\x3c\xbe\x93\x9e\xc9"
"\x3b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 33,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x4e\xf6\xfa\x13\xde\x43\x63\x4c"
"\xe2\x04\x3e\xe4\x85\x14\xb6\x3f"
"\xb1\x8f\x4c\xdb\x41\xa2\x14\x99"
"\xf5\x53\x0f\x73\x86\x7e\x97\xa1"
"\x4b\x56\x5b\x94\xce\xcd\x74\xcd"
"\x75\xc4\x53\x01\x89\x45\x59",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 47,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xa4\x9a\xb7\xfd\xa0\xd4\xd6\x47"
"\x95\xf4\x58\x38\x14\x83\x27\x01"
"\x4c\xed\x32\x2c\xf7\xd6\x31\xf7"
"\x38\x1b\x2c\xc9\xb6\x31\xce\xaa"
"\xa5\x3c\x1a\x18\x5c\xce\xb9\xdf"
"\x51\x52\x77\xf2\x5e\x85\x80\x41",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 48,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xfb\xd4\x83\x71\x9e\x63\xad\x60"
"\xb9\xf9\xeb\x34\x52\x49\xcf\xb7",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 15,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x0c\xaf\x2e\x96\xf6\x97\x08\x71"
"\x7d\x3a\x84\xc4\x44\x57\x77\x7e",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xc7\x87\x09\x3b\xc7\x19\x74\x22"
"\x22\xa5\x67\x10\xb2\x36\xb3\x45",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 17,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x02\xc6\x3b\x46\x65\xb2\xef\x91"
"\x31\xf0\x45\x48\x8a\x2a\xed\xe4",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 31,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x20\x85\xa8\xd0\x91\x48\x85\xf3"
"\x5a\x16\xc0\x57\x68\x47\xdd\xcb",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 32,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x6a\xf8\x8d\x9c\x42\x75\x35\x79"
"\xc1\x96\xbd\x31\x6e\x69\x1b\x50",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54",
.assoc = "\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x01\x24\xb1\xba\xf6\xd3\xdf\x83"
"\x70\x45\xe3\x2a\x9d\x5c\x63\x98"
"\x39",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x6d\x92\x42\x61\xa7\x58\x37",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 15,
.input = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 15,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x18\x78\xc2\x6e\xe1\xf7\xe6\x8a"
"\xca\x0e\x62\x00\xa8\x21\xb5\x21"
"\x3d\x36\xdb\xf7\xcc\x31\x94\x9c"
"\x98\xbd\x71\x7a\xef\xa4\xfa",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 16,
.input = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xea\xd1\x81\x75\xb4\x13\x1d\x86"
"\xd4\x17\x26\xe5\xd6\x89\x39\x04"
"\xa9\x6c\xca\xac\x40\x73\xb2\x4c"
"\x9c\xb9\x0e\x79\x4c\x40\x65\xc6",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 17,
.input = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xf4\xb2\x84\xd1\x81\xfa\x98\x1c"
"\x38\x2d\x69\x90\x1c\x71\x38\x98"
"\x9f\xe1\x19\x3b\x63\x91\xaf\x6e"
"\x4b\x07\x2c\xac\x53\xc5\xd5\xfe"
"\x93",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 33,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 31,
.input = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xa0\xe7\x0a\x60\xe7\xb8\x8a\xdb"
"\x94\xd3\x93\xf2\x41\x86\x16\xdd"
"\x4c\xe8\xe7\xe0\x62\x48\x89\x40"
"\xc0\x49\x9b\x63\x32\xec\x8b\xdb"
"\xdc\xa6\xea\x2c\xc2\x7f\xf5\x04"
"\xcb\xe5\x47\xbb\xa7\xd1\x9d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 47,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 32,
.input = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x62\xdc\x2d\x68\x2d\x71\xbb\x33"
"\x13\xdf\xc0\x46\xf6\x61\x94\xa7"
"\x60\xd3\xd4\xca\xd9\xbe\x82\xf3"
"\xf1\x5b\xa0\xfa\x15\xba\xda\xea"
"\x87\x68\x47\x08\x5d\xdd\x83\xb0"
"\x60\xf4\x93\x20\xdf\x34\x8f\xea",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 48,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x9d",
.alen = 33,
.input = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 65,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x84\xc5\x21\xab\xe1\xeb\xbb\x6d"
"\xaa\x2a\xaf\xeb\x3b\x3b\x69\xe7"
"\x2c\x47\xef\x9d\xb7\x53\x36\xb7"
"\xb6\xf5\xe5\xa8\xc9\x9e\x02\xd7"
"\x83\x88\xc2\xbd\x2f\xf9\x10\xc0"
"\xf5\xa1\x6e\xd3\x97\x64\x82\xa3"
"\xfb\xda\x2c\xb1\x94\xa1\x58\x32"
"\xe8\xd4\x39\xfc\x9e\x26\xf9\xf1"
"\x61\xe6\xae\x07\xf2\xe0\xa7\x44"
"\x96\x28\x3b\xee\x6b\xc6\x16\x31"
"\x3f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 81,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
.alen = 65,
.input = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
"\x2f",
.ilen = 33,
.result = "\x8f\x23\x47\xfb\xf2\xac\x23\x83"
"\x77\x09\xac\x74\xef\xd2\x56\xae"
"\x20\x7b\x7b\xca\x45\x8e\xc8\xc2"
"\x50\xbd\xc7\x44\x1c\x54\x98\xd8"
"\x1f\xd0\x9a\x79\xaa\xf9\xe1\xb3"
"\xb4\x98\x5a\x9b\xe4\x4d\xbf\x4e"
"\x39",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 49,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 16,
.input = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x42\xc3\x58\xfb\x29\xe2\x4a\x56"
"\xf1\xf5\xe1\x51\x55\x4b\x0a\x45"
"\x46\xb5\x8d\xac\xb6\x34\xd8\x8b"
"\xde\x20\x59\x77\xc1\x74\x90",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 16,
.input = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xb2\xfb\xf6\x97\x69\x7a\xe9\xec"
"\xe2\x94\xa1\x8b\xa0\x2b\x60\x72"
"\x1d\x04\xdd\x6a\xef\x46\x8f\x68"
"\xe9\xe0\x17\x45\x70\x12",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 30,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x47\xda\x54\x42\x51\x72\xc4\x8b"
"\xf5\x57\x0f\x2f\x49\x0e\x11\x3b"
"\x78\x93\xec\xfc\xf4\xff\xe1\x2d",
.rlen = 24,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* AEGIS-128 test vectors - generated via reference implementation from
* SUPERCOP (https://bench.cr.yp.to/supercop.html):
*
* https://bench.cr.yp.to/supercop/supercop-20170228.tar.xz
* (see crypto_aead/aegis128/)
*/
static const struct aead_testvec aegis128_dec_tv_template[] = {
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x07\xa5\x11\xf2\x9d\x40\xb8\x6d"
"\xda\xb8\x12\x34\x4c\x53\xd9\x72",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x9e\x78\x52\xae\xcb\x9e\xe4\xd3"
"\x9a\xd7\x5d\xd7\xaa\x9a\xe9\x5a"
"\xcc",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x79",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 1,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xc3\x80\x83\x04\x5f\xaa\x61\xc7"
"\xca\xdd\x6f\xac\x85\x08\xb5\x35"
"\x2b\xc2\x3e\x0b\x1b\x39\x37\x2b"
"\x7a\x21\x16\xb3\xe6\x67\x66",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 15,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x23\x25\x30\xe5\x6a\xb6\x36\x7d"
"\x38\xfd\x3a\xd2\xc2\x58\xa9\x11"
"\x1e\xa8\x30\x9c\x16\xa4\xdb\x65"
"\x51\x10\x16\x27\x70\x9b\x64\x29",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2a\x8d\x56\x91\xc6\xf3\x56\xa5"
"\x1f\xf0\x89\x2e\x13\xad\xe6\xf6"
"\x46\x80\xb1\x0e\x18\x30\x40\x97"
"\x03\xdf\x64\x3c\xbe\x93\x9e\xc9"
"\x3b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 33,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4e\xf6\xfa\x13\xde\x43\x63\x4c"
"\xe2\x04\x3e\xe4\x85\x14\xb6\x3f"
"\xb1\x8f\x4c\xdb\x41\xa2\x14\x99"
"\xf5\x53\x0f\x73\x86\x7e\x97\xa1"
"\x4b\x56\x5b\x94\xce\xcd\x74\xcd"
"\x75\xc4\x53\x01\x89\x45\x59",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 47,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa4\x9a\xb7\xfd\xa0\xd4\xd6\x47"
"\x95\xf4\x58\x38\x14\x83\x27\x01"
"\x4c\xed\x32\x2c\xf7\xd6\x31\xf7"
"\x38\x1b\x2c\xc9\xb6\x31\xce\xaa"
"\xa5\x3c\x1a\x18\x5c\xce\xb9\xdf"
"\x51\x52\x77\xf2\x5e\x85\x80\x41",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 48,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xfb\xd4\x83\x71\x9e\x63\xad\x60"
"\xb9\xf9\xeb\x34\x52\x49\xcf\xb7",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 15,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x0c\xaf\x2e\x96\xf6\x97\x08\x71"
"\x7d\x3a\x84\xc4\x44\x57\x77\x7e",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xc7\x87\x09\x3b\xc7\x19\x74\x22"
"\x22\xa5\x67\x10\xb2\x36\xb3\x45",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x02\xc6\x3b\x46\x65\xb2\xef\x91"
"\x31\xf0\x45\x48\x8a\x2a\xed\xe4",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x20\x85\xa8\xd0\x91\x48\x85\xf3"
"\x5a\x16\xc0\x57\x68\x47\xdd\xcb",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x6a\xf8\x8d\x9c\x42\x75\x35\x79"
"\xc1\x96\xbd\x31\x6e\x69\x1b\x50",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54",
.assoc = "\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x01\x24\xb1\xba\xf6\xd3\xdf\x83"
"\x70\x45\xe3\x2a\x9d\x5c\x63\x98"
"\x39",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x4f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 1,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 15,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x18\x78\xc2\x6e\xe1\xf7\xe6\x8a"
"\xca\x0e\x62\x00\xa8\x21\xb5\x21"
"\x3d\x36\xdb\xf7\xcc\x31\x94\x9c"
"\x98\xbd\x71\x7a\xef\xa4\xfa",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 15,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xea\xd1\x81\x75\xb4\x13\x1d\x86"
"\xd4\x17\x26\xe5\xd6\x89\x39\x04"
"\xa9\x6c\xca\xac\x40\x73\xb2\x4c"
"\x9c\xb9\x0e\x79\x4c\x40\x65\xc6",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 17,
.input = "\xf4\xb2\x84\xd1\x81\xfa\x98\x1c"
"\x38\x2d\x69\x90\x1c\x71\x38\x98"
"\x9f\xe1\x19\x3b\x63\x91\xaf\x6e"
"\x4b\x07\x2c\xac\x53\xc5\xd5\xfe"
"\x93",
.ilen = 33,
.result = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 31,
.input = "\xa0\xe7\x0a\x60\xe7\xb8\x8a\xdb"
"\x94\xd3\x93\xf2\x41\x86\x16\xdd"
"\x4c\xe8\xe7\xe0\x62\x48\x89\x40"
"\xc0\x49\x9b\x63\x32\xec\x8b\xdb"
"\xdc\xa6\xea\x2c\xc2\x7f\xf5\x04"
"\xcb\xe5\x47\xbb\xa7\xd1\x9d",
.ilen = 47,
.result = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 32,
.input = "\x62\xdc\x2d\x68\x2d\x71\xbb\x33"
"\x13\xdf\xc0\x46\xf6\x61\x94\xa7"
"\x60\xd3\xd4\xca\xd9\xbe\x82\xf3"
"\xf1\x5b\xa0\xfa\x15\xba\xda\xea"
"\x87\x68\x47\x08\x5d\xdd\x83\xb0"
"\x60\xf4\x93\x20\xdf\x34\x8f\xea",
.ilen = 48,
.result = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x9d",
.alen = 33,
.input = "\x84\xc5\x21\xab\xe1\xeb\xbb\x6d"
"\xaa\x2a\xaf\xeb\x3b\x3b\x69\xe7"
"\x2c\x47\xef\x9d\xb7\x53\x36\xb7"
"\xb6\xf5\xe5\xa8\xc9\x9e\x02\xd7"
"\x83\x88\xc2\xbd\x2f\xf9\x10\xc0"
"\xf5\xa1\x6e\xd3\x97\x64\x82\xa3"
"\xfb\xda\x2c\xb1\x94\xa1\x58\x32"
"\xe8\xd4\x39\xfc\x9e\x26\xf9\xf1"
"\x61\xe6\xae\x07\xf2\xe0\xa7\x44"
"\x96\x28\x3b\xee\x6b\xc6\x16\x31"
"\x3f",
.ilen = 81,
.result = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 65,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
.alen = 65,
.input = "\x8f\x23\x47\xfb\xf2\xac\x23\x83"
"\x77\x09\xac\x74\xef\xd2\x56\xae"
"\x20\x7b\x7b\xca\x45\x8e\xc8\xc2"
"\x50\xbd\xc7\x44\x1c\x54\x98\xd8"
"\x1f\xd0\x9a\x79\xaa\xf9\xe1\xb3"
"\xb4\x98\x5a\x9b\xe4\x4d\xbf\x4e"
"\x39",
.ilen = 49,
.result = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
"\x2f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 33,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x42\xc3\x58\xfb\x29\xe2\x4a\x56"
"\xf1\xf5\xe1\x51\x55\x4b\x0a\x45"
"\x46\xb5\x8d\xac\xb6\x34\xd8\x8b"
"\xde\x20\x59\x77\xc1\x74\x90",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xb2\xfb\xf6\x97\x69\x7a\xe9\xec"
"\xe2\x94\xa1\x8b\xa0\x2b\x60\x72"
"\x1d\x04\xdd\x6a\xef\x46\x8f\x68"
"\xe9\xe0\x17\x45\x70\x12",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 30,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x47\xda\x54\x42\x51\x72\xc4\x8b"
"\xf5\x57\x0f\x2f\x49\x0e\x11\x3b"
"\x78\x93\xec\xfc\xf4\xff\xe1\x2d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 24,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
.rlen = 16,
},
};
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* AEGIS-128L test vectors - generated via reference implementation from
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
* SUPERCOP (https://bench.cr.yp.to/supercop.html):
*
* https://bench.cr.yp.to/supercop/supercop-20170228.tar.xz
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* (see crypto_aead/aegis128l/)
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec aegis128l_enc_tv_template[] = {
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x30\x4f\xf3\xe9\xb1\xfa\x81\xa6"
"\x20\x72\x78\xdd\x93\xc8\x57\xef",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x79",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xa9\x24\xa0\xb6\x2d\xdd\x29\xdb"
"\x40\xb3\x71\xc5\x22\x58\x31\x77"
"\x6d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
.ilen = 15,
.result = "\xbb\x0a\x53\xc4\xaa\x7e\xa4\x03"
"\x2b\xee\x62\x99\x7b\x98\x13\x1f"
"\xe0\x76\x4c\x2e\x53\x99\x4f\xbe"
"\xe1\xa8\x04\x7f\xe1\x71\xbe",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ilen = 16,
.result = "\x66\xdf\x6e\x71\xc0\x6e\xa4\x4c"
"\x9d\xb7\x8c\x9a\xdb\x1f\xd2\x2e"
"\x23\xb6\xa4\xfb\xd3\x86\xdd\xbb"
"\xde\x54\x9b\xf5\x92\x8b\x93\xc5",
.rlen = 32,
}, {
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
.ilen = 17,
.result = "\x4f\xc3\x69\xb6\xd3\xa4\x64\x8b"
"\x71\xc3\x8a\x91\x22\x4f\x1b\xd2"
"\x33\x6d\x86\xbc\xf8\x2f\x06\xf9"
"\x82\x64\xc7\x72\x00\x30\xfc\xf0"
"\xf8",
.rlen = 33,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
.ilen = 31,
.result = "\xe3\x93\x15\xae\x5f\x9d\x3c\xb5"
"\xd6\x9d\xee\xee\xcf\xaa\xaf\xe1"
"\x45\x10\x96\xe0\xbf\x55\x0f\x4c"
"\x1a\xfd\xf4\xda\x4e\x10\xde\xc9"
"\x0e\x6f\xc7\x3c\x49\x94\x41\xfc"
"\x59\x28\x88\x3c\x79\x10\x6b",
.rlen = 47,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
.ilen = 32,
.result = "\x1c\x8e\x22\x34\xfd\xab\xe6\x0d"
"\x1c\x9f\x06\x54\x8b\x0b\xb4\x40"
"\xde\x11\x59\x3e\xfd\x74\xf6\x42"
"\x97\x17\xf7\x24\xb6\x7e\xc4\xc6"
"\x06\xa3\x94\xda\x3d\x7f\x55\x0a"
"\x92\x07\x2f\xa6\xf3\x6b\x2c\xfc",
.rlen = 48,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xa0\x2a\xb4\x9a\x91\x00\x15\xb8"
"\x0f\x9a\x15\x60\x0e\x9b\x13\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
.alen = 15,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x4c\x26\xad\x9c\x14\xfd\x9c\x8c"
"\x84\xfb\x26\xfb\xd5\xca\x62\x39",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
.alen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x45\x85\x0e\x0f\xf4\xae\x96\xa1"
"\x99\x4d\x6d\xb4\x67\x32\xb0\x3a",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
.alen = 17,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x33\xb1\x42\x97\x8e\x16\x7b\x63"
"\x06\xba\x5b\xcb\xae\x6d\x8b\x56",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
.alen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xda\x44\x08\x8c\x2a\xa5\x07\x35"
"\x0b\x54\x4e\x6d\xe3\xfd\xc4\x5f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
.alen = 32,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.input = "",
.ilen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x1b\xb1\xf1\xa8\x9e\xc2\xb2\x88"
"\x40\x7f\x7b\x19\x7a\x52\x8c\xf0",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54",
.assoc = "\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x6e\xc8\xfb\x15\x9d\x98\x49\xc9"
"\xa0\x98\x09\x85\xbe\x56\x8e\x79"
"\xf4",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.alen = 15,
.input = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
.ilen = 15,
.result = "\x99\x2e\x84\x50\x64\x5c\xab\x29"
"\x20\xba\xb9\x2f\x62\x3a\xce\x2a"
"\x75\x25\x3b\xe3\x40\xe0\x1d\xfc"
"\x20\x63\x0b\x49\x7e\x97\x08",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.alen = 16,
.input = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.ilen = 16,
.result = "\xd9\x8e\xfd\x50\x8f\x02\x9f\xee"
"\x78\x08\x12\xec\x09\xaf\x53\x14"
"\x90\x3e\x3d\x76\xad\x71\x21\x08"
"\x77\xe5\x4b\x15\xc2\xe6\xbc\xdb",
.rlen = 32,
}, {
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x05",
.alen = 17,
.input = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
.ilen = 17,
.result = "\xf3\xe7\x95\x86\xcf\x34\x95\x96"
"\x17\xfe\x1b\xae\x1b\x31\xf2\x1a"
"\xbd\xbc\xc9\x4e\x11\x29\x09\x5c"
"\x05\xd3\xb4\x2e\x4a\x74\x59\x49"
"\x7d",
.rlen = 33,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x68\x28\x73\x40\x9f\x96\x4a",
.alen = 31,
.input = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
.ilen = 31,
.result = "\x06\x96\xb2\xbf\x63\xf4\x1e\x24"
"\x0d\x19\x15\x61\x65\x3b\x06\x26"
"\x71\xe8\x7e\x16\xdb\x96\x01\x01"
"\x52\xcd\x49\x5b\x07\x33\x4e\xe7"
"\xaa\x91\xf5\xd5\xc6\xfe\x41\xb5"
"\xed\x90\xce\xb9\xcd\xcc\xa1",
.rlen = 47,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
.alen = 32,
.input = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
.ilen = 32,
.result = "\xf9\xd7\xee\x17\xfd\x24\xcd\xf1"
"\xbc\x0f\x35\x97\x97\x0c\x4b\x18"
"\xce\x58\xc8\x3b\xd4\x85\x93\x79"
"\xcc\x9c\xea\xc1\x73\x13\x0b\x4c"
"\xcc\x6f\x28\xf8\xa4\x4e\xb8\x56"
"\x64\x4e\x47\xce\xb2\xb4\x92\xb4",
.rlen = 48,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x9d",
.alen = 33,
.input = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
.ilen = 65,
.result = "\x58\xfa\x3a\x3d\xd9\x88\x63\xe8"
"\xc5\x78\x50\x8b\x4a\xc9\xdf\x7f"
"\x4b\xfa\xc8\x2e\x67\x43\xf3\x63"
"\x42\x8e\x99\x5a\x9c\x0b\x84\x77"
"\xbc\x46\x76\x48\x82\xc7\x57\x96"
"\xe1\x65\xd1\xed\x1d\xdd\x80\x24"
"\xa6\x4d\xa9\xf1\x53\x8b\x5e\x0e"
"\x26\xb9\xcc\x37\xe5\x43\xe1\x5a"
"\x8a\xd6\x8c\x5a\xe4\x95\xd1\x8d"
"\xf7\x33\x64\xc1\xd3\xf2\xfc\x35"
"\x01",
.rlen = 81,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.klen = 16,
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 65,
.input = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
"\x2f",
.ilen = 33,
.result = "\x4c\xa9\xac\x71\xed\x10\xa6\x24"
"\xb7\xa7\xdf\x8b\xf5\xc2\x41\xcb"
"\x05\xc9\xd6\x97\xb6\x10\x7f\x17"
"\xc2\xc0\x93\xcf\xe0\x94\xfd\x99"
"\xf2\x62\x25\x28\x01\x23\x6f\x8b"
"\x04\x52\xbc\xb0\x3e\x66\x52\x90"
"\x9f",
.rlen = 49,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.alen = 16,
.input = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
.ilen = 16,
.result = "\x6d\xed\x04\x7a\x2f\x0c\x30\xa5"
"\x96\xe6\x97\xe4\x10\xeb\x40\x95"
"\xc5\x9a\xdf\x31\xd5\xa5\xa6\xec"
"\x05\xa8\x31\x50\x11\x19\x44",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.alen = 16,
.input = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
.ilen = 16,
.result = "\x30\x95\x7d\xea\xdc\x62\xc0\x88"
"\xa1\xe3\x8d\x8c\xac\x04\x10\xa7"
"\xfa\xfa\x07\xbd\xa0\xf0\x36\xeb"
"\x21\x93\x2e\x31\x84\x83",
.rlen = 30,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.klen = 16,
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.alen = 16,
.input = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ilen = 16,
.result = "\x93\xcd\xee\xd4\xcb\x9d\x8d\x16"
"\x63\x0d\x43\xd5\x49\xca\xa8\x85"
"\x49\xc0\xae\x13\xbc\x26\x1d\x4b",
.rlen = 24,
},
};
static const struct aead_testvec aegis128l_dec_tv_template[] = {
{
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x30\x4f\xf3\xe9\xb1\xfa\x81\xa6"
"\x20\x72\x78\xdd\x93\xc8\x57\xef",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
.klen = 16,
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa9\x24\xa0\xb6\x2d\xdd\x29\xdb"
"\x40\xb3\x71\xc5\x22\x58\x31\x77"
"\x6d",
.ilen = 17,
.result = "\x79",
.rlen = 1,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
.klen = 16,
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xbb\x0a\x53\xc4\xaa\x7e\xa4\x03"
"\x2b\xee\x62\x99\x7b\x98\x13\x1f"
"\xe0\x76\x4c\x2e\x53\x99\x4f\xbe"
"\xe1\xa8\x04\x7f\xe1\x71\xbe",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
.rlen = 15,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
.klen = 16,
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x66\xdf\x6e\x71\xc0\x6e\xa4\x4c"
"\x9d\xb7\x8c\x9a\xdb\x1f\xd2\x2e"
"\x23\xb6\xa4\xfb\xd3\x86\xdd\xbb"
"\xde\x54\x9b\xf5\x92\x8b\x93\xc5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
.klen = 16,
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4f\xc3\x69\xb6\xd3\xa4\x64\x8b"
"\x71\xc3\x8a\x91\x22\x4f\x1b\xd2"
"\x33\x6d\x86\xbc\xf8\x2f\x06\xf9"
"\x82\x64\xc7\x72\x00\x30\xfc\xf0"
"\xf8",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 33,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
.rlen = 17,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
.klen = 16,
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xe3\x93\x15\xae\x5f\x9d\x3c\xb5"
"\xd6\x9d\xee\xee\xcf\xaa\xaf\xe1"
"\x45\x10\x96\xe0\xbf\x55\x0f\x4c"
"\x1a\xfd\xf4\xda\x4e\x10\xde\xc9"
"\x0e\x6f\xc7\x3c\x49\x94\x41\xfc"
"\x59\x28\x88\x3c\x79\x10\x6b",
.ilen = 47,
.result = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
.klen = 16,
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x1c\x8e\x22\x34\xfd\xab\xe6\x0d"
"\x1c\x9f\x06\x54\x8b\x0b\xb4\x40"
"\xde\x11\x59\x3e\xfd\x74\xf6\x42"
"\x97\x17\xf7\x24\xb6\x7e\xc4\xc6"
"\x06\xa3\x94\xda\x3d\x7f\x55\x0a"
"\x92\x07\x2f\xa6\xf3\x6b\x2c\xfc",
.ilen = 48,
.result = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
.rlen = 32,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
.klen = 16,
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa0\x2a\xb4\x9a\x91\x00\x15\xb8"
"\x0f\x9a\x15\x60\x0e\x9b\x13\x8f",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
.klen = 16,
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
.alen = 15,
.input = "\x4c\x26\xad\x9c\x14\xfd\x9c\x8c"
"\x84\xfb\x26\xfb\xd5\xca\x62\x39",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
.klen = 16,
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
.alen = 16,
.input = "\x45\x85\x0e\x0f\xf4\xae\x96\xa1"
"\x99\x4d\x6d\xb4\x67\x32\xb0\x3a",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
.klen = 16,
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
.alen = 17,
.input = "\x33\xb1\x42\x97\x8e\x16\x7b\x63"
"\x06\xba\x5b\xcb\xae\x6d\x8b\x56",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
.klen = 16,
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xda\x44\x08\x8c\x2a\xa5\x07\x35"
"\x0b\x54\x4e\x6d\xe3\xfd\xc4\x5f",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
.klen = 16,
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x1b\xb1\xf1\xa8\x9e\xc2\xb2\x88"
"\x40\x7f\x7b\x19\x7a\x52\x8c\xf0",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
.klen = 16,
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54",
.assoc = "\x40",
.alen = 1,
.input = "\x6e\xc8\xfb\x15\x9d\x98\x49\xc9"
"\xa0\x98\x09\x85\xbe\x56\x8e\x79"
"\xf4",
.ilen = 17,
.result = "\x4f",
.rlen = 1,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.klen = 16,
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.alen = 15,
.input = "\x99\x2e\x84\x50\x64\x5c\xab\x29"
"\x20\xba\xb9\x2f\x62\x3a\xce\x2a"
"\x75\x25\x3b\xe3\x40\xe0\x1d\xfc"
"\x20\x63\x0b\x49\x7e\x97\x08",
.ilen = 31,
.result = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
.rlen = 15,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
.klen = 16,
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.alen = 16,
.input = "\xd9\x8e\xfd\x50\x8f\x02\x9f\xee"
"\x78\x08\x12\xec\x09\xaf\x53\x14"
"\x90\x3e\x3d\x76\xad\x71\x21\x08"
"\x77\xe5\x4b\x15\xc2\xe6\xbc\xdb",
.ilen = 32,
.result = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
.klen = 16,
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
.alen = 17,
.input = "\xf3\xe7\x95\x86\xcf\x34\x95\x96"
"\x17\xfe\x1b\xae\x1b\x31\xf2\x1a"
"\xbd\xbc\xc9\x4e\x11\x29\x09\x5c"
"\x05\xd3\xb4\x2e\x4a\x74\x59\x49"
"\x7d",
.ilen = 33,
.result = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 17,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
.klen = 16,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 31,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x06\x96\xb2\xbf\x63\xf4\x1e\x24"
"\x0d\x19\x15\x61\x65\x3b\x06\x26"
"\x71\xe8\x7e\x16\xdb\x96\x01\x01"
"\x52\xcd\x49\x5b\x07\x33\x4e\xe7"
"\xaa\x91\xf5\xd5\xc6\xfe\x41\xb5"
"\xed\x90\xce\xb9\xcd\xcc\xa1",
.ilen = 47,
.result = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
.klen = 16,
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xf9\xd7\xee\x17\xfd\x24\xcd\xf1"
"\xbc\x0f\x35\x97\x97\x0c\x4b\x18"
"\xce\x58\xc8\x3b\xd4\x85\x93\x79"
"\xcc\x9c\xea\xc1\x73\x13\x0b\x4c"
"\xcc\x6f\x28\xf8\xa4\x4e\xb8\x56"
"\x64\x4e\x47\xce\xb2\xb4\x92\xb4",
.ilen = 48,
.result = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
.rlen = 32,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
.klen = 16,
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 33,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x58\xfa\x3a\x3d\xd9\x88\x63\xe8"
"\xc5\x78\x50\x8b\x4a\xc9\xdf\x7f"
"\x4b\xfa\xc8\x2e\x67\x43\xf3\x63"
"\x42\x8e\x99\x5a\x9c\x0b\x84\x77"
"\xbc\x46\x76\x48\x82\xc7\x57\x96"
"\xe1\x65\xd1\xed\x1d\xdd\x80\x24"
"\xa6\x4d\xa9\xf1\x53\x8b\x5e\x0e"
"\x26\xb9\xcc\x37\xe5\x43\xe1\x5a"
"\x8a\xd6\x8c\x5a\xe4\x95\xd1\x8d"
"\xf7\x33\x64\xc1\xd3\xf2\xfc\x35"
"\x01",
.ilen = 81,
.result = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
.rlen = 65,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
.klen = 16,
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 65,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4c\xa9\xac\x71\xed\x10\xa6\x24"
"\xb7\xa7\xdf\x8b\xf5\xc2\x41\xcb"
"\x05\xc9\xd6\x97\xb6\x10\x7f\x17"
"\xc2\xc0\x93\xcf\xe0\x94\xfd\x99"
"\xf2\x62\x25\x28\x01\x23\x6f\x8b"
"\x04\x52\xbc\xb0\x3e\x66\x52\x90"
"\x9f",
.ilen = 49,
.result = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x2f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.rlen = 33,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.klen = 16,
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.alen = 16,
.input = "\x6d\xed\x04\x7a\x2f\x0c\x30\xa5"
"\x96\xe6\x97\xe4\x10\xeb\x40\x95"
"\xc5\x9a\xdf\x31\xd5\xa5\xa6\xec"
"\x05\xa8\x31\x50\x11\x19\x44",
.ilen = 31,
.result = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
.klen = 16,
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.alen = 16,
.input = "\x30\x95\x7d\xea\xdc\x62\xc0\x88"
"\xa1\xe3\x8d\x8c\xac\x04\x10\xa7"
"\xfa\xfa\x07\xbd\xa0\xf0\x36\xeb"
"\x21\x93\x2e\x31\x84\x83",
.ilen = 30,
.result = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.klen = 16,
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.alen = 16,
.input = "\x93\xcd\xee\xd4\xcb\x9d\x8d\x16"
"\x63\x0d\x43\xd5\x49\xca\xa8\x85"
"\x49\xc0\xae\x13\xbc\x26\x1d\x4b",
.ilen = 24,
.result = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* AEGIS-256 test vectors - generated via reference implementation from
* SUPERCOP (https://bench.cr.yp.to/supercop.html):
*
* https://bench.cr.yp.to/supercop/supercop-20170228.tar.xz
* (see crypto_aead/aegis256/)
*/
static const struct aead_testvec aegis256_enc_tv_template[] = {
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81"
"\xca\xb0\x82\x21\x41\xa8\xe0\x06"
"\x30\x0b\x37\xf6\xb6\x17\xe7\xb5",
.klen = 32,
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03"
"\x95\x61\x05\x42\x82\x50\xc0\x0c"
"\x60\x16\x6f\xec\x6d\x2f\xcf\x6b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "",
.ilen = 0,
.result = "\xd5\x65\x3a\xa9\x03\x51\xd7\xaa"
"\xfa\x4b\xd8\xa2\x41\x9b\xc1\xb2",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87"
"\xf4\x72\x8e\xa5\x46\x48\x62\x20"
"\xf1\x38\x16\xce\x90\x76\x87\x8c",
.klen = 32,
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09"
"\xbf\x23\x11\xc6\x87\xf0\x42\x26"
"\x22\x44\x4e\xc4\x47\x8e\x6e\x41",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x79",
.ilen = 1,
.result = "\x84\xa2\x8f\xad\xdb\x8d\x2c\x16"
"\x9e\x89\xd9\x06\xa6\xa8\x14\x29"
"\x8b",
.rlen = 17,
}, {
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e"
"\x1f\x35\x9a\x29\x4b\xe8\xe4\x39"
"\xb3\x66\xf5\xa6\x6a\xd5\x26\x62",
.klen = 32,
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f"
"\xe9\xe5\x1d\x4a\x8c\x90\xc4\x40"
"\xe3\x71\x2d\x9c\x21\xed\x0e\x18",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
.ilen = 15,
.result = "\x09\x94\x1f\xa6\x13\xc3\x74\x75"
"\x17\xad\x8a\x0e\xd8\x66\x9a\x28"
"\xd7\x30\x66\x09\x2a\xdc\xfa\x2a"
"\x9f\x3b\xd7\xdd\x66\xd1\x2b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94"
"\x49\xf7\xa5\xad\x50\x88\x66\x53"
"\x74\x94\xd4\x7f\x44\x34\xc5\x39",
.klen = 32,
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15"
"\x14\xa8\x28\xce\x92\x30\x46\x59"
"\xa4\x9f\x0b\x75\xfb\x4c\xad\xee",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.ilen = 16,
.result = "\x8a\x46\xa2\x22\x8c\x03\xab\x6f"
"\x54\x63\x4e\x7f\xc9\x8e\xfa\x70"
"\x7b\xe5\x8d\x78\xbc\xe9\xb6\xa1"
"\x29\x17\xc8\x3b\x52\xa4\x98\x72",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a"
"\x74\xb9\xb1\x32\x55\x28\xe8\x6d"
"\x35\xc1\xb3\x57\x1f\x93\x64\x0f",
.klen = 32,
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c"
"\x3e\x6a\x34\x53\x97\xd0\xc8\x73"
"\x66\xcd\xea\x4d\xd5\xab\x4c\xc5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
.ilen = 17,
.result = "\x71\x6b\x37\x0b\x02\x61\x28\x12"
"\x83\xab\x66\x90\x84\xc7\xd1\xc5"
"\xb2\x7a\xb4\x7b\xb4\xfe\x02\xb2"
"\xc0\x00\x39\x13\xb5\x51\x68\x44"
"\xad",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 33,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0"
"\x9e\x7c\xbc\xb6\x5b\xc8\x6a\x86"
"\xf7\xef\x91\x30\xf9\xf2\x04\xe6",
.klen = 32,
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22"
"\x69\x2c\x3f\xd7\x9c\x70\x4a\x8d"
"\x27\xfa\xc9\x26\xaf\x0a\xeb\x9c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
.ilen = 31,
.result = "\xaf\xa4\x34\x0d\x59\xe6\x1c\x2f"
"\x06\x3b\x52\x18\x49\x75\x1b\xf0"
"\x53\x09\x72\x7b\x45\x79\xe0\xbe"
"\x89\x85\x23\x15\xb8\x79\x07\x4c"
"\x53\x7a\x15\x37\x0a\xee\xb7\xfb"
"\xc4\x1f\x12\x27\xcf\x77\x90",
.rlen = 47,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6"
"\xc8\x3e\xc8\x3a\x60\x68\xec\xa0"
"\xb8\x1c\x70\x08\xd3\x51\xa3\xbd",
.klen = 32,
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28"
"\x93\xef\x4b\x5b\xa1\x10\xcc\xa6"
"\xe8\x28\xa8\xfe\x89\x69\x8b\x72",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
.ilen = 32,
.result = "\xe2\xc9\x0b\x33\x31\x02\xb3\xb4"
"\x33\xfe\xeb\xa8\xb7\x9b\xb2\xd7"
"\xeb\x0f\x05\x2b\xba\xb3\xca\xef"
"\xf6\xd1\xb6\xc0\xb9\x9b\x85\xc5"
"\xbf\x7a\x3e\xcc\x31\x76\x09\x80"
"\x32\x5d\xbb\xe8\x38\x0e\x77\xd3",
.rlen = 48,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad"
"\xf3\x00\xd4\xbf\x65\x08\x6e\xb9"
"\x7a\x4a\x4f\xe0\xad\xb0\x42\x93",
.klen = 32,
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e"
"\xbd\xb1\x57\xe0\xa6\xb0\x4e\xc0"
"\xaa\x55\x87\xd6\x63\xc8\x2a\x49",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "",
.ilen = 0,
.result = "\x96\x43\x30\xca\x6c\x4f\xd7\x12"
"\xba\xd9\xb3\x18\x86\xdf\xc3\x52",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3"
"\x1d\xc3\xdf\x43\x6a\xa8\xf0\xd3"
"\x3b\x77\x2e\xb9\x87\x0f\xe1\x6a",
.klen = 32,
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34"
"\xe8\x73\x62\x64\xab\x50\xd0\xda"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x6b\x83\x66\xaf\x3e\x27\xc9\x1f",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
.alen = 15,
.input = "",
.ilen = 0,
.result = "\x2f\xab\x45\xe2\xa7\x46\xc5\x83"
"\x11\x9f\xb0\x74\xee\xc7\x03\xdd",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9"
"\x47\x85\xeb\xc7\x6f\x48\x72\xed"
"\xfc\xa5\x0d\x91\x61\x6e\x81\x40",
.klen = 32,
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b"
"\x12\x35\x6e\xe8\xb0\xf0\x52\xf3"
"\x2d\xb0\x45\x87\x18\x86\x68\xf6",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
.alen = 16,
.input = "",
.ilen = 0,
.result = "\x16\x44\x73\x33\x5d\xf2\xb9\x04"
"\x6b\x79\x98\xef\xdb\xd5\xc5\xf1",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf"
"\x72\x47\xf6\x4b\x74\xe8\xf4\x06"
"\xbe\xd3\xec\x6a\x3b\xcd\x20\x17",
.klen = 32,
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c\xf8\x79\x6c\xb6\x90\xd4\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xee\xde\x23\x60\xf2\xe5\x08\xcc",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
.alen = 17,
.input = "",
.ilen = 0,
.result = "\xa4\x9b\xb8\x47\xc0\xed\x7a\x45"
"\x98\x54\x8c\xed\x3d\x17\xf0\xdd",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6"
"\x9c\x0a\x02\xd0\x79\x88\x76\x20"
"\x7f\x00\xca\x42\x15\x2c\xbf\xed",
.klen = 32,
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xaf\x0b\x02\x38\xcc\x44\xa7\xa3",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
.alen = 31,
.input = "",
.ilen = 0,
.result = "\x20\x24\xe2\x33\x5c\x60\xc9\xf0"
"\xa4\x96\x2f\x0d\x53\xc2\xf8\xfc",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc"
"\xc6\xcc\x0e\x54\x7f\x28\xf8\x3a"
"\x40\x2e\xa9\x1a\xf0\x8b\x5e\xc4",
.klen = 32,
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
.alen = 32,
.input = "",
.ilen = 0,
.result = "\x6f\x4a\xb9\xe0\xff\x51\xa3\xf1"
"\xd2\x64\x3e\x66\x6a\xb2\x03\xc0",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2"
"\xf1\x8e\x19\xd8\x84\xc8\x7a\x53"
"\x02\x5b\x88\xf3\xca\xea\xfe\x9b",
.klen = 32,
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54"
"\xbb\x3f\x9c\xf9\xc5\x70\x5a\x5a"
"\x32\x67\xc0\xe9\x80\x02\xe5\x50",
.assoc = "\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x4f",
.ilen = 1,
.result = "\x2c\xfb\xad\x7e\xbe\xa0\x9a\x5b"
"\x7a\x3f\x81\xf7\xfc\x1b\x79\x83"
"\xc7",
.rlen = 17,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8"
"\x1b\x50\x25\x5d\x89\x68\xfc\x6d"
"\xc3\x89\x67\xcb\xa4\x49\x9d\x71",
.klen = 32,
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a"
"\xe6\x01\xa8\x7e\xca\x10\xdc\x73"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xf4\x94\x9f\xc1\x5a\x61\x85\x27",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.alen = 15,
.input = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
.ilen = 15,
.result = "\x1f\x7f\xca\x3c\x2b\xe7\x27\xba"
"\x7e\x98\x83\x02\x34\x23\xf7\x94"
"\xde\x35\xe6\x1d\x14\x18\xe5\x38"
"\x14\x80\x6a\xa7\x1b\xae\x1d",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 31,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf"
"\x46\x13\x31\xe1\x8e\x08\x7e\x87"
"\x85\xb6\x46\xa3\x7e\xa8\x3c\x48",
.klen = 32,
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60"
"\x10\xc3\xb3\x02\xcf\xb0\x5e\x8d"
"\xb5\xc2\x7e\x9a\x35\xc0\x24\xfd",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.alen = 16,
.input = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.ilen = 16,
.result = "\x05\x86\x9e\xd7\x2b\xa3\x97\x01"
"\xbe\x28\x98\x10\x6f\xe9\x61\x32"
"\x96\xbb\xb1\x2e\x8f\x0c\x44\xb9"
"\x46\x2d\x55\xe3\x42\x67\xf2\xaf",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5"
"\x70\xd5\x3c\x65\x93\xa8\x00\xa0"
"\x46\xe4\x25\x7c\x58\x08\xdb\x1e",
.klen = 32,
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b\x86\xbf\x86\xd4\x50\xe0\xa7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x76\xef\x5c\x72\x0f\x1f\xc3\xd4",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x05",
.alen = 17,
.input = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
.ilen = 17,
.result = "\x9c\xe0\x06\x7b\x86\xcf\x2e\xd8"
"\x45\x65\x1b\x72\x9b\xaa\xa3\x1e"
"\x87\x9d\x26\xdf\xff\x81\x11\xd2"
"\x47\x41\xb9\x24\xc1\x8a\xa3\x8b"
"\x55",
.rlen = 33,
}, {
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb"
"\x9a\x97\x48\xe9\x98\x48\x82\xba"
"\x07\x11\x04\x54\x32\x67\x7b\xf5",
.klen = 32,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x38\x1d\x3b\x4a\xe9\x7e\x62\xaa",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x68\x28\x73\x40\x9f\x96\x4a",
.alen = 31,
.input = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
.ilen = 31,
.result = "\xa0\xc8\xde\x83\x0d\xc3\x4e\xd5"
"\x69\x7f\x7a\xdd\x8c\x46\xda\xba"
"\x0a\x5c\x0e\x7f\xac\xee\x02\xd2"
"\xe5\x4b\x0a\xba\xb8\xa4\x7b\x66"
"\xde\xae\xdb\xc2\xc0\x0b\xf7\x2b"
"\xdf\xb8\xea\xd8\xa9\x38\xed",
.rlen = 47,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1"
"\xc5\x5a\x53\x6e\x9d\xe8\x04\xd4"
"\xc9\x3f\xe2\x2d\x0c\xc6\x1a\xcb",
.klen = 32,
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
.alen = 32,
.input = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
.ilen = 32,
.result = "\xd3\x68\x14\x70\x3c\x01\x43\x86"
"\x02\xab\xbe\x75\xaa\xe7\xf5\x53"
"\x5c\x05\xbd\x9b\x19\xbb\x2a\x61"
"\x8f\x69\x05\x75\x8e\xca\x60\x0c"
"\x5b\xa2\x48\x61\x32\x74\x11\x2b"
"\xf6\xcf\x06\x78\x6f\x78\x1a\x4a",
.rlen = 48,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7"
"\xef\x1c\x5f\xf2\xa3\x88\x86\xed"
"\x8a\x6d\xc1\x05\xe7\x25\xb9\xa2",
.klen = 32,
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x9d",
.alen = 33,
.input = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
.ilen = 65,
.result = "\x07\x0a\x35\xb0\x82\x03\x5a\xd2"
"\x15\x3a\x6c\x72\x83\x9b\xb1\x75"
"\xea\xf2\xfc\xff\xc6\xf1\x13\xa4"
"\x1a\x93\x33\x79\x97\x82\x81\xc0"
"\x96\xc2\x00\xab\x39\xae\xa1\x62"
"\x53\xa3\x86\xc9\x07\x8c\xaf\x22"
"\x47\x31\x29\xca\x4a\x95\xf5\xd5"
"\x20\x63\x5a\x54\x80\x2c\x4a\x63"
"\xfb\x18\x73\x31\x4f\x08\x21\x5d"
"\x20\xe9\xc3\x7e\xea\x25\x77\x3a"
"\x65",
.rlen = 81,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe"
"\x19\xde\x6b\x76\xa8\x28\x08\x07"
"\x4b\x9a\xa0\xdd\xc1\x84\x58\x79",
.klen = 32,
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.alen = 65,
.input = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
"\x2f",
.ilen = 33,
.result = "\x33\xc1\xda\xfa\x15\x21\x07\x8e"
"\x93\x68\xea\x64\x7b\x3d\x4b\x6b"
"\x71\x5e\x5e\x6b\x92\xaa\x65\xc2"
"\x7a\x2a\xc1\xa9\x0a\xa1\x24\x81"
"\x26\x3a\x5a\x09\xe8\xce\x73\x72"
"\xde\x7b\x58\x9e\x85\xb9\xa4\x28"
"\xda",
.rlen = 49,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04"
"\x44\xa1\x76\xfb\xad\xc8\x8a\x21"
"\x0d\xc8\x7f\xb6\x9b\xe3\xf8\x4f",
.klen = 32,
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85"
"\x0e\x51\xf9\x1c\xee\x70\x6a\x27"
"\x3d\xd3\xb7\xac\x51\xfa\xdf\x05",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.alen = 16,
.input = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ilen = 16,
.result = "\x3e\xf8\x86\x3d\x39\xf8\x96\x02"
"\x0f\xdf\xc9\x6e\x37\x1e\x57\x99"
"\x07\x2a\x1a\xac\xd1\xda\xfd\x3b"
"\xc7\xff\xbd\xbc\x85\x09\x0b",
.rlen = 31,
}, {
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a"
"\x6e\x63\x82\x7f\xb2\x68\x0c\x3a"
"\xce\xf5\x5e\x8e\x75\x42\x97\x26",
.klen = 32,
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c"
"\x39\x14\x05\xa0\xf3\x10\xec\x41"
"\xff\x01\x95\x84\x2b\x59\x7f\xdb",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.alen = 16,
.input = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ilen = 16,
.result = "\x2f\xc4\xd8\x0d\xa6\x07\xef\x2e"
"\x6c\xd9\x84\x63\x70\x97\x61\x37"
"\x08\x2f\x16\x90\x9e\x62\x30\x0d"
"\x62\xd5\xc8\xf0\x46\x1a",
.rlen = 30,
}, {
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10"
"\x98\x25\x8d\x03\xb7\x08\x8e\x54"
"\x90\x23\x3d\x67\x4f\xa1\x36\xfc",
.klen = 32,
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\xd5\x07\x58\x59\x72\xd7\xde\x92"
"\x63\xd6\x10\x24\xf8\xb0\x6e\x5a"
"\xc0\x2e\x74\x5d\x06\xb8\x1e\xb2",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.alen = 16,
.input = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ilen = 16,
.result = "\xce\xf3\x17\x87\x49\xc2\x00\x46"
"\xc6\x12\x5c\x8f\x81\x38\xaa\x55"
"\xf8\x67\x75\xf1\x75\xe3\x2a\x24",
.rlen = 24,
},
};
static const struct aead_testvec aegis256_dec_tv_template[] = {
{
.key = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81"
"\xca\xb0\x82\x21\x41\xa8\xe0\x06"
"\x30\x0b\x37\xf6\xb6\x17\xe7\xb5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x1e\x92\x1c\xcf\x88\x3d\x54\x0d"
"\x40\x6d\x59\x48\xfc\x92\x61\x03"
"\x95\x61\x05\x42\x82\x50\xc0\x0c"
"\x60\x16\x6f\xec\x6d\x2f\xcf\x6b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xd5\x65\x3a\xa9\x03\x51\xd7\xaa"
"\xfa\x4b\xd8\xa2\x41\x9b\xc1\xb2",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87"
"\xf4\x72\x8e\xa5\x46\x48\x62\x20"
"\xf1\x38\x16\xce\x90\x76\x87\x8c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x5a\xb7\x56\x6e\x98\xb9\xfd\x29"
"\xc1\x47\x0b\xda\xf6\xb6\x23\x09"
"\xbf\x23\x11\xc6\x87\xf0\x42\x26"
"\x22\x44\x4e\xc4\x47\x8e\x6e\x41",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x84\xa2\x8f\xad\xdb\x8d\x2c\x16"
"\x9e\x89\xd9\x06\xa6\xa8\x14\x29"
"\x8b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x79",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 1,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e"
"\x1f\x35\x9a\x29\x4b\xe8\xe4\x39"
"\xb3\x66\xf5\xa6\x6a\xd5\x26\x62",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x97\xdb\x90\x0e\xa8\x35\xa5\x45"
"\x42\x21\xbd\x6b\xf0\xda\xe6\x0f"
"\xe9\xe5\x1d\x4a\x8c\x90\xc4\x40"
"\xe3\x71\x2d\x9c\x21\xed\x0e\x18",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x09\x94\x1f\xa6\x13\xc3\x74\x75"
"\x17\xad\x8a\x0e\xd8\x66\x9a\x28"
"\xd7\x30\x66\x09\x2a\xdc\xfa\x2a"
"\x9f\x3b\xd7\xdd\x66\xd1\x2b",
.ilen = 31,
.result = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47",
.rlen = 15,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94"
"\x49\xf7\xa5\xad\x50\x88\x66\x53"
"\x74\x94\xd4\x7f\x44\x34\xc5\x39",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xd3\x00\xc9\xad\xb8\xb0\x4e\x61"
"\xc3\xfb\x6f\xfd\xea\xff\xa9\x15"
"\x14\xa8\x28\xce\x92\x30\x46\x59"
"\xa4\x9f\x0b\x75\xfb\x4c\xad\xee",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x8a\x46\xa2\x22\x8c\x03\xab\x6f"
"\x54\x63\x4e\x7f\xc9\x8e\xfa\x70"
"\x7b\xe5\x8d\x78\xbc\xe9\xb6\xa1"
"\x29\x17\xc8\x3b\x52\xa4\x98\x72",
.ilen = 32,
.result = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a"
"\x74\xb9\xb1\x32\x55\x28\xe8\x6d"
"\x35\xc1\xb3\x57\x1f\x93\x64\x0f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x10\x25\x03\x4c\xc8\x2c\xf7\x7d"
"\x44\xd5\x21\x8e\xe4\x23\x6b\x1c"
"\x3e\x6a\x34\x53\x97\xd0\xc8\x73"
"\x66\xcd\xea\x4d\xd5\xab\x4c\xc5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x71\x6b\x37\x0b\x02\x61\x28\x12"
"\x83\xab\x66\x90\x84\xc7\xd1\xc5"
"\xb2\x7a\xb4\x7b\xb4\xfe\x02\xb2"
"\xc0\x00\x39\x13\xb5\x51\x68\x44"
"\xad",
.ilen = 33,
.result = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f"
"\xd3",
.rlen = 17,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0"
"\x9e\x7c\xbc\xb6\x5b\xc8\x6a\x86"
"\xf7\xef\x91\x30\xf9\xf2\x04\xe6",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x4c\x49\x3d\xec\xd8\xa8\xa0\x98"
"\xc5\xb0\xd3\x1f\xde\x48\x2e\x22"
"\x69\x2c\x3f\xd7\x9c\x70\x4a\x8d"
"\x27\xfa\xc9\x26\xaf\x0a\xeb\x9c",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xaf\xa4\x34\x0d\x59\xe6\x1c\x2f"
"\x06\x3b\x52\x18\x49\x75\x1b\xf0"
"\x53\x09\x72\x7b\x45\x79\xe0\xbe"
"\x89\x85\x23\x15\xb8\x79\x07\x4c"
"\x53\x7a\x15\x37\x0a\xee\xb7\xfb"
"\xc4\x1f\x12\x27\xcf\x77\x90",
.ilen = 47,
.result = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25"
"\xfe\x8d\x45\x19\x1e\xc0\x0b\x99"
"\x88\x11\x39\x12\x1c\x3a\xbb",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6"
"\xc8\x3e\xc8\x3a\x60\x68\xec\xa0"
"\xb8\x1c\x70\x08\xd3\x51\xa3\xbd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x89\x6e\x77\x8b\xe8\x23\x49\xb4"
"\x45\x8a\x85\xb1\xd8\x6c\xf1\x28"
"\x93\xef\x4b\x5b\xa1\x10\xcc\xa6"
"\xe8\x28\xa8\xfe\x89\x69\x8b\x72",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.assoc = "",
.alen = 0,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xe2\xc9\x0b\x33\x31\x02\xb3\xb4"
"\x33\xfe\xeb\xa8\xb7\x9b\xb2\xd7"
"\xeb\x0f\x05\x2b\xba\xb3\xca\xef"
"\xf6\xd1\xb6\xc0\xb9\x9b\x85\xc5"
"\xbf\x7a\x3e\xcc\x31\x76\x09\x80"
"\x32\x5d\xbb\xe8\x38\x0e\x77\xd3",
.ilen = 48,
.result = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b"
"\x28\x50\x51\x9d\x24\x60\x8d\xb3"
"\x49\x3e\x17\xea\xf6\x99\x5a\xdd",
.rlen = 32,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad"
"\xf3\x00\xd4\xbf\x65\x08\x6e\xb9"
"\x7a\x4a\x4f\xe0\xad\xb0\x42\x93",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xc5\x93\xb0\x2a\xf8\x9f\xf1\xd0"
"\xc6\x64\x37\x42\xd2\x90\xb3\x2e"
"\xbd\xb1\x57\xe0\xa6\xb0\x4e\xc0"
"\xaa\x55\x87\xd6\x63\xc8\x2a\x49",
.assoc = "\xd5",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x96\x43\x30\xca\x6c\x4f\xd7\x12"
"\xba\xd9\xb3\x18\x86\xdf\xc3\x52",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3"
"\x1d\xc3\xdf\x43\x6a\xa8\xf0\xd3"
"\x3b\x77\x2e\xb9\x87\x0f\xe1\x6a",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34"
"\xe8\x73\x62\x64\xab\x50\xd0\xda"
"\x6b\x83\x66\xaf\x3e\x27\xc9\x1f",
.assoc = "\x11\x81\x78\x32\x4d\xb9\x44\x73"
"\x68\x75\x16\xf8\xcb\x7e\xa7",
.alen = 15,
.input = "\x2f\xab\x45\xe2\xa7\x46\xc5\x83"
"\x11\x9f\xb0\x74\xee\xc7\x03\xdd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9"
"\x47\x85\xeb\xc7\x6f\x48\x72\xed"
"\xfc\xa5\x0d\x91\x61\x6e\x81\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b"
"\x12\x35\x6e\xe8\xb0\xf0\x52\xf3"
"\x2d\xb0\x45\x87\x18\x86\x68\xf6",
.assoc = "\x4e\xa5\xb2\xd1\x5d\x35\xed\x8f"
"\xe8\x4f\xc8\x89\xc5\xa2\x69\xbc",
.alen = 16,
.input = "\x16\x44\x73\x33\x5d\xf2\xb9\x04"
"\x6b\x79\x98\xef\xdb\xd5\xc5\xf1",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf"
"\x72\x47\xf6\x4b\x74\xe8\xf4\x06"
"\xbe\xd3\xec\x6a\x3b\xcd\x20\x17",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c\xf8\x79\x6c\xb6\x90\xd4\x0d"
"\xee\xde\x23\x60\xf2\xe5\x08\xcc",
.assoc = "\x8a\xca\xec\x70\x6d\xb1\x96\xab"
"\x69\x29\x7a\x1b\xbf\xc7\x2c\xc2"
"\x07",
.alen = 17,
.input = "\xa4\x9b\xb8\x47\xc0\xed\x7a\x45"
"\x98\x54\x8c\xed\x3d\x17\xf0\xdd",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6"
"\x9c\x0a\x02\xd0\x79\x88\x76\x20"
"\x7f\x00\xca\x42\x15\x2c\xbf\xed",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
"\xaf\x0b\x02\x38\xcc\x44\xa7\xa3",
.assoc = "\xc7\xef\x26\x10\x7d\x2c\x3f\xc6"
"\xea\x03\x2c\xac\xb9\xeb\xef\xc9"
"\x31\x6b\x08\x12\xfc\xd8\x37\x2d"
"\xe0\x17\x3a\x2e\x83\x5c\x8f",
.alen = 31,
.input = "\x20\x24\xe2\x33\x5c\x60\xc9\xf0"
"\xa4\x96\x2f\x0d\x53\xc2\xf8\xfc",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc"
"\xc6\xcc\x0e\x54\x7f\x28\xf8\x3a"
"\x40\x2e\xa9\x1a\xf0\x8b\x5e\xc4",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a",
.assoc = "\x03\x14\x5f\xaf\x8d\xa8\xe7\xe2"
"\x6b\xde\xde\x3e\xb3\x10\xb1\xcf"
"\x5c\x2d\x14\x96\x01\x78\xb9\x47"
"\xa1\x44\x19\x06\x5d\xbb\x2e\x2f",
.alen = 32,
.input = "\x6f\x4a\xb9\xe0\xff\x51\xa3\xf1"
"\xd2\x64\x3e\x66\x6a\xb2\x03\xc0",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 16,
.result = "",
.rlen = 0,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2"
"\xf1\x8e\x19\xd8\x84\xc8\x7a\x53"
"\x02\x5b\x88\xf3\xca\xea\xfe\x9b",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x31\x6f\x0b\xe6\x59\x85\xe6\x77"
"\xcc\x81\x63\xab\xae\x6b\x43\x54"
"\xbb\x3f\x9c\xf9\xc5\x70\x5a\x5a"
"\x32\x67\xc0\xe9\x80\x02\xe5\x50",
.assoc = "\x40",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\x2c\xfb\xad\x7e\xbe\xa0\x9a\x5b"
"\x7a\x3f\x81\xf7\xfc\x1b\x79\x83"
"\xc7",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x4f",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 1,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8"
"\x1b\x50\x25\x5d\x89\x68\xfc\x6d"
"\xc3\x89\x67\xcb\xa4\x49\x9d\x71",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a"
"\xe6\x01\xa8\x7e\xca\x10\xdc\x73"
"\xf4\x94\x9f\xc1\x5a\x61\x85\x27",
.assoc = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.alen = 15,
.input = "\x1f\x7f\xca\x3c\x2b\xe7\x27\xba"
"\x7e\x98\x83\x02\x34\x23\xf7\x94"
"\xde\x35\xe6\x1d\x14\x18\xe5\x38"
"\x14\x80\x6a\xa7\x1b\xae\x1d",
.ilen = 31,
.result = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67",
.rlen = 15,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf"
"\x46\x13\x31\xe1\x8e\x08\x7e\x87"
"\x85\xb6\x46\xa3\x7e\xa8\x3c\x48",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60"
"\x10\xc3\xb3\x02\xcf\xb0\x5e\x8d"
"\xb5\xc2\x7e\x9a\x35\xc0\x24\xfd",
.assoc = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.alen = 16,
.input = "\x05\x86\x9e\xd7\x2b\xa3\x97\x01"
"\xbe\x28\x98\x10\x6f\xe9\x61\x32"
"\x96\xbb\xb1\x2e\x8f\x0c\x44\xb9"
"\x46\x2d\x55\xe3\x42\x67\xf2\xaf",
.ilen = 32,
.result = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.rlen = 16,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5"
"\x70\xd5\x3c\x65\x93\xa8\x00\xa0"
"\x46\xe4\x25\x7c\x58\x08\xdb\x1e",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b\x86\xbf\x86\xd4\x50\xe0\xa7"
"\x76\xef\x5c\x72\x0f\x1f\xc3\xd4",
.assoc = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
.alen = 17,
.input = "\x9c\xe0\x06\x7b\x86\xcf\x2e\xd8"
"\x45\x65\x1b\x72\x9b\xaa\xa3\x1e"
"\x87\x9d\x26\xdf\xff\x81\x11\xd2"
"\x47\x41\xb9\x24\xc1\x8a\xa3\x8b"
"\x55",
.ilen = 33,
.result = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69"
"\xd0",
.rlen = 17,
}, {
.key = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb"
"\x9a\x97\x48\xe9\x98\x48\x82\xba"
"\x07\x11\x04\x54\x32\x67\x7b\xf5",
.klen = 32,
.iv = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
"\x38\x1d\x3b\x4a\xe9\x7e\x62\xaa",
.assoc = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
.alen = 31,
.input = "\xa0\xc8\xde\x83\x0d\xc3\x4e\xd5"
"\x69\x7f\x7a\xdd\x8c\x46\xda\xba"
"\x0a\x5c\x0e\x7f\xac\xee\x02\xd2"
"\xe5\x4b\x0a\xba\xb8\xa4\x7b\x66"
"\xde\xae\xdb\xc2\xc0\x0b\xf7\x2b"
"\xdf\xb8\xea\xd8\xa9\x38\xed",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.ilen = 47,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70"
"\xfa\xa9\xd0\x4d\x5c\x40\x23\xcd"
"\x98\x34\xab\x37\x56\xae\x32",
.rlen = 31,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1"
"\xc5\x5a\x53\x6e\x9d\xe8\x04\xd4"
"\xc9\x3f\xe2\x2d\x0c\xc6\x1a\xcb",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81",
.assoc = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.alen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.input = "\xd3\x68\x14\x70\x3c\x01\x43\x86"
"\x02\xab\xbe\x75\xaa\xe7\xf5\x53"
"\x5c\x05\xbd\x9b\x19\xbb\x2a\x61"
"\x8f\x69\x05\x75\x8e\xca\x60\x0c"
"\x5b\xa2\x48\x61\x32\x74\x11\x2b"
"\xf6\xcf\x06\x78\x6f\x78\x1a\x4a",
.ilen = 48,
.result = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76"
"\x24\x6b\xdc\xd1\x61\xe0\xa5\xe7"
"\x5a\x61\x8a\x0f\x30\x0d\xd1\xec",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7"
"\xef\x1c\x5f\xf2\xa3\x88\x86\xed"
"\x8a\x6d\xc1\x05\xe7\x25\xb9\xa2",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58",
.assoc = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d",
.alen = 33,
.input = "\x07\x0a\x35\xb0\x82\x03\x5a\xd2"
"\x15\x3a\x6c\x72\x83\x9b\xb1\x75"
"\xea\xf2\xfc\xff\xc6\xf1\x13\xa4"
"\x1a\x93\x33\x79\x97\x82\x81\xc0"
"\x96\xc2\x00\xab\x39\xae\xa1\x62"
"\x53\xa3\x86\xc9\x07\x8c\xaf\x22"
"\x47\x31\x29\xca\x4a\x95\xf5\xd5"
"\x20\x63\x5a\x54\x80\x2c\x4a\x63"
"\xfb\x18\x73\x31\x4f\x08\x21\x5d"
"\x20\xe9\xc3\x7e\xea\x25\x77\x3a"
"\x65",
.ilen = 81,
.result = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c"
"\x4f\x2e\xe8\x55\x66\x80\x27\x00"
"\x1b\x8f\x68\xe7\x0a\x6c\x71\xc3"
"\x21\x78\x55\x9d\x9c\x65\x7b\xcd"
"\x0a\x34\x97\xff\x47\x37\xb0\x2a"
"\x80\x0d\x19\x98\x33\xa9\x7a\xe3"
"\x2e\x4c\xc6\xf3\x8c\x88\x42\x01"
"\xbd",
.rlen = 65,
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe"
"\x19\xde\x6b\x76\xa8\x28\x08\x07"
"\x4b\x9a\xa0\xdd\xc1\x84\x58\x79",
crypto: testmgr - Add test vectors for MORUS This patch adds test vectors for MORUS-640 and MORUS-1280. The test vectors were generated using the reference implementation from SUPERCOP (see code comments for more details). Signed-off-by: Ondrej Mosnacek <omosnacek@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-11 06:19:10 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e",
.assoc = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
.alen = 65,
.input = "\x33\xc1\xda\xfa\x15\x21\x07\x8e"
"\x93\x68\xea\x64\x7b\x3d\x4b\x6b"
"\x71\x5e\x5e\x6b\x92\xaa\x65\xc2"
"\x7a\x2a\xc1\xa9\x0a\xa1\x24\x81"
"\x26\x3a\x5a\x09\xe8\xce\x73\x72"
"\xde\x7b\x58\x9e\x85\xb9\xa4\x28"
"\xda",
.ilen = 49,
.result = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82"
"\x79\xf0\xf3\xd9\x6c\x20\xa9\x1a"
"\xdc\xbc\x47\xc0\xe4\xcb\x10\x99"
"\x2f",
.rlen = 33,
}, {
.key = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04"
"\x44\xa1\x76\xfb\xad\xc8\x8a\x21"
"\x0d\xc8\x7f\xb6\x9b\xe3\xf8\x4f",
.klen = 32,
.iv = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85"
"\x0e\x51\xf9\x1c\xee\x70\x6a\x27"
"\x3d\xd3\xb7\xac\x51\xfa\xdf\x05",
.assoc = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.alen = 16,
.input = "\x3e\xf8\x86\x3d\x39\xf8\x96\x02"
"\x0f\xdf\xc9\x6e\x37\x1e\x57\x99"
"\x07\x2a\x1a\xac\xd1\xda\xfd\x3b"
"\xc7\xff\xbd\xbc\x85\x09\x0b",
.ilen = 31,
.result = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
crypto: keywrap - add testmgr support The testmanager code for symmetric ciphers is extended to allow verification of the IV after a cipher operation. In addition, test vectors for kw(aes) for encryption and decryption are added. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-09-21 12:59:56 -06:00
.rlen = 16,
crypto: testmgr - add extra kw(aes) encryption test vector One "kw(aes)" decryption test vector doesn't exactly match an encryption test vector with input and result swapped. In preparation for removing the decryption test vectors, add this test vector to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:28 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a"
"\x6e\x63\x82\x7f\xb2\x68\x0c\x3a"
"\xce\xf5\x5e\x8e\x75\x42\x97\x26",
crypto: testmgr - add extra kw(aes) encryption test vector One "kw(aes)" decryption test vector doesn't exactly match an encryption test vector with input and result swapped. In preparation for removing the decryption test vectors, add this test vector to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:28 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c"
"\x39\x14\x05\xa0\xf3\x10\xec\x41"
"\xff\x01\x95\x84\x2b\x59\x7f\xdb",
.assoc = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.alen = 16,
.input = "\x2f\xc4\xd8\x0d\xa6\x07\xef\x2e"
"\x6c\xd9\x84\x63\x70\x97\x61\x37"
"\x08\x2f\x16\x90\x9e\x62\x30\x0d"
"\x62\xd5\xc8\xf0\x46\x1a",
.ilen = 30,
.result = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
crypto: testmgr - add extra kw(aes) encryption test vector One "kw(aes)" decryption test vector doesn't exactly match an encryption test vector with input and result swapped. In preparation for removing the decryption test vectors, add this test vector to the encryption test vectors, so we don't lose any test coverage. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:28 -06:00
.rlen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}, {
.key = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10"
"\x98\x25\x8d\x03\xb7\x08\x8e\x54"
"\x90\x23\x3d\x67\x4f\xa1\x36\xfc",
crypto: keywrap - add testmgr support The testmanager code for symmetric ciphers is extended to allow verification of the IV after a cipher operation. In addition, test vectors for kw(aes) for encryption and decryption are added. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-09-21 12:59:56 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92"
"\x63\xd6\x10\x24\xf8\xb0\x6e\x5a"
"\xc0\x2e\x74\x5d\x06\xb8\x1e\xb2",
.assoc = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.alen = 16,
.input = "\xce\xf3\x17\x87\x49\xc2\x00\x46"
"\xc6\x12\x5c\x8f\x81\x38\xaa\x55"
"\xf8\x67\x75\xf1\x75\xe3\x2a\x24",
.ilen = 24,
.result = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
crypto: keywrap - add testmgr support The testmanager code for symmetric ciphers is extended to allow verification of the IV after a cipher operation. In addition, test vectors for kw(aes) for encryption and decryption are added. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-09-21 12:59:56 -06:00
.rlen = 16,
},
};
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* MORUS-640 test vectors - generated via reference implementation from
* SUPERCOP (https://bench.cr.yp.to/supercop.html):
*
* https://bench.cr.yp.to/supercop/supercop-20170228.tar.xz
* (see crypto_aead/morus640128v2/)
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec morus640_enc_tv_template[] = {
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.iv = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
.assoc = "",
.alen = 0,
.input = "",
.ilen = 0,
.result = "\x89\x62\x7d\xf3\x07\x9d\x52\x05"
"\x53\xc3\x04\x60\x93\xb4\x37\x9a",
.rlen = 16,
}, {
.key = "\x3c\x24\x39\x9f\x10\x7b\xa8\x1b"
"\x80\xda\xb2\x91\xf9\x24\xc2\x06",
.klen = 16,
.iv = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
.assoc = "",
.alen = 0,
.input = "\x69",
.ilen = 1,
.result = "\xa8\x8d\xe4\x90\xb5\x50\x8f\x78"
"\xb6\x10\x9a\x59\x5f\x61\x37\x70"
"\x09",
.rlen = 17,
}, {
.key = "\x79\x49\x73\x3e\x20\xf7\x51\x37"
"\x01\xb4\x64\x22\xf3\x48\x85\x0c",
.klen = 16,
.iv = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
.assoc = "",
.alen = 0,
.input = "\xa6\xa4\x1e\x76\xec\xd4\x50\xcc"
"\x62\x58\xe9\x8f\xef\xa4\x17",
.ilen = 15,
.result = "\x76\xdd\xb9\x05\x3d\xce\x61\x38"
"\xf3\xef\xf7\xe5\xd7\xfd\x70\xa5"
"\xcf\x9d\x64\xb8\x0a\x9f\xfd\x8b"
"\xd4\x6e\xfe\xd9\xc8\x63\x4b",
.rlen = 31,
}, {
.key = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47\x12",
.klen = 16,
.iv = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
.assoc = "",
.alen = 0,
.input = "\xe2\xc9\x58\x15\xfc\x4f\xf8\xe8"
"\xe3\x32\x9b\x21\xe9\xc8\xd9\x97",
.ilen = 16,
.result = "\xdc\x72\xe8\x14\xfb\x63\xad\x72"
"\x1f\x57\x9a\x1f\x88\x81\xdb\xd6"
"\xc1\x91\x9d\xb9\x25\xc4\x99\x4c"
"\x97\xcd\x8a\x0c\x9d\x68\x00\x1c",
.rlen = 32,
}, {
.key = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.klen = 16,
.iv = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
.assoc = "",
.alen = 0,
.input = "\x1f\xee\x92\xb4\x0c\xcb\xa1\x04"
"\x64\x0c\x4d\xb2\xe3\xec\x9c\x9d"
"\x09",
.ilen = 17,
.result = "\x6b\x4f\x3b\x90\x9a\xa2\xb3\x82"
"\x0a\xb8\x55\xee\xeb\x73\x4d\x7f"
"\x54\x11\x3a\x8a\x31\xa3\xb5\xf2"
"\xcd\x49\xdb\xf3\xee\x26\xbd\xa2"
"\x0d",
.rlen = 33,
}, {
.key = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f",
.klen = 16,
.iv = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
.assoc = "",
.alen = 0,
.input = "\x5c\x13\xcb\x54\x1c\x47\x4a\x1f"
"\xe5\xe6\xff\x44\xdd\x11\x5f\xa3"
"\x33\xdd\xc2\xf8\xdd\x18\x2b\x93"
"\x57\x05\x01\x1c\x66\x22\xd3",
.ilen = 31,
.result = "\x59\xd1\x0f\x6b\xee\x27\x84\x92"
"\xb7\xa9\xb5\xdd\x02\xa4\x12\xa5"
"\x50\x32\xb4\x9a\x2e\x35\x83\x55"
"\x36\x12\x12\xed\xa3\x31\xc5\x30"
"\xa7\xe2\x4a\x6d\x05\x59\x43\x91"
"\x75\xfa\x6c\x17\xc6\x73\xca",
.rlen = 47,
}, {
.key = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25",
.klen = 16,
.iv = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
.assoc = "",
.alen = 0,
.input = "\x98\x37\x05\xf3\x2c\xc2\xf3\x3b"
"\x66\xc0\xb1\xd5\xd7\x35\x21\xaa"
"\x5d\x9f\xce\x7c\xe2\xb8\xad\xad"
"\x19\x33\xe0\xf4\x40\x81\x72\x28",
.ilen = 32,
.result = "\xdb\x49\x68\x0f\x91\x5b\x21\xb1"
"\xcf\x50\xb2\x4c\x32\xe1\xa6\x69"
"\xc0\xfb\x44\x1f\xa0\x9a\xeb\x39"
"\x1b\xde\x68\x38\xcc\x27\x52\xc5"
"\xf6\x3e\x74\xea\x66\x5b\x5f\x0c"
"\x65\x9e\x58\xe6\x52\xa2\xfe\x59",
.rlen = 48,
}, {
.key = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
.assoc = "\xc5",
.alen = 1,
.input = "",
.ilen = 0,
.result = "\x56\xe7\x24\x52\xdd\x95\x60\x5b"
"\x09\x48\x39\x69\x9c\xb3\x62\x46",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe4\x25\xcd\xfa\x80\xdd\x46\xde"
"\x07\xd1\x90\x8b\xcf\x23\x15\x31",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
.assoc = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76",
.alen = 15,
.input = "",
.ilen = 0,
.result = "\xdd\xfa\x6c\x1f\x5d\x86\x87\x01"
"\x13\xe5\x73\x46\x46\xf2\x5c\xe1",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x20\x4a\x07\x99\x91\x58\xee\xfa"
"\x88\xab\x42\x1c\xc9\x47\xd7\x38",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
.assoc = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
.alen = 16,
.input = "",
.ilen = 0,
.result = "\xa6\x1b\xb9\xd7\x5e\x3c\xcf\xac"
"\xa9\x21\x45\x0b\x16\x52\xf7\xe1",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5d\x6f\x41\x39\xa1\xd4\x97\x16"
"\x09\x85\xf4\xae\xc3\x6b\x9a\x3e",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
.assoc = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c",
.alen = 17,
.input = "",
.ilen = 0,
.result = "\x15\xff\xde\x3b\x34\xfc\xf6\xf9"
"\xbb\xa8\x62\xad\x0a\xf5\x48\x60",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x99\x93\x7a\xd8\xb1\x50\x40\x31"
"\x8a\x60\xa6\x3f\xbd\x90\x5d\x44",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
.assoc = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
"\xaf\x0b\x02\x38\xcc\x44\xa7",
.alen = 31,
.input = "",
.ilen = 0,
.result = "\xd2\x9d\xf8\x3b\xd7\x84\xe9\x2d"
"\x4b\xef\x75\x16\x0a\x99\xae\x6b",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}, {
.key = "\xd6\xb8\xb4\x77\xc1\xcb\xe9\x4d"
"\x0a\x3a\x58\xd1\xb7\xb4\x1f\x4a",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
.assoc = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a",
.alen = 32,
.input = "",
.ilen = 0,
.result = "\xe4\x8d\xa7\xa7\x45\xc1\x31\x4f"
"\xce\xfb\xaf\xd6\xc2\xe6\xee\xc0",
crypto: testmgr - Add ansi_cprng test vectors Add ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode), aka 'ansi_cprng' test vectors, taken from Appendix B.2.9 and B.2.10 of the NIST RNGVS document, found here: http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf Successfully tested against both the cryptodev-2.6 tree and a Red Hat Enterprise Linux 5.4 kernel, via 'modprobe tcrypt mode=150'. The selection of 150 was semi-arbitrary, didn't seem like it should go any place in particular, so I started a new range for rng tests. Signed-off-by: Jarod Wilson <jarod@redhat.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-05-04 05:46:29 -06:00
.rlen = 16,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x12\xdd\xee\x17\xd1\x47\x92\x69"
"\x8b\x14\x0a\x62\xb1\xd9\xe2\x50",
.klen = 16,
.iv = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
.assoc = "\x31",
.alen = 1,
.input = "\x40",
.ilen = 1,
.result = "\xe2\x67\x38\x4f\xb9\xad\x7d\x38"
"\x01\xfe\x84\x14\x85\xf8\xd1\xe3"
"\x22",
.rlen = 17,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x4f\x01\x27\xb6\xe1\xc3\x3a\x85"
"\x0c\xee\xbc\xf4\xab\xfd\xa5\x57",
.klen = 16,
.iv = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.assoc = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06",
.alen = 15,
.input = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.ilen = 15,
.result = "\x77\x32\x61\xeb\xb4\x33\x29\x92"
"\x29\x95\xc5\x8e\x85\x76\xab\xfc"
"\x07\x95\xa7\x44\x74\xf7\x22\xff"
"\xd8\xd8\x36\x3d\x8a\x7f\x9e",
.rlen = 31,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67\x5d",
.klen = 16,
.iv = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
.assoc = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
.alen = 16,
.input = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.ilen = 16,
.result = "\xd8\xfd\x44\x45\xf6\x42\x12\x38"
"\xf2\x0b\xea\x4f\x9e\x11\x61\x07"
"\x48\x67\x98\x18\x9b\xd0\x0c\x59"
"\x67\xa4\x11\xb3\x2b\xd6\xc1\x70",
.rlen = 32,
}, {
.key = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.klen = 16,
.iv = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
.assoc = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b",
.alen = 17,
.input = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
.ilen = 17,
.result = "\xb1\xab\x53\x4e\xc7\x40\x16\xb6"
"\x71\x3a\x00\x9f\x41\x88\xb0\xb2"
"\x71\x83\x85\x5f\xc8\x79\x0a\x99"
"\x99\xdc\x89\x1c\x88\xd2\x3e\xf9"
"\x83",
.rlen = 33,
}, {
.key = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69",
.klen = 16,
.iv = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
.assoc = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
"\x38\x1d\x3b\x4a\xe9\x7e\x62",
.alen = 31,
.input = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
.ilen = 31,
.result = "\x29\xc4\xf0\x03\xc1\x86\xdf\x06"
"\x5c\x7b\xef\x64\x87\x00\xd1\x37"
"\xa7\x08\xbc\x7f\x8f\x41\x54\xd0"
"\x3e\xf1\xc3\xa2\x96\x84\xdd\x2a"
"\x2d\x21\x30\xf9\x02\xdb\x06\x0c"
"\xf1\x5a\x66\x69\xe0\xca\x83",
.rlen = 47,
}, {
.key = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70",
.klen = 16,
.iv = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
.assoc = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81",
.alen = 32,
.input = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
.ilen = 32,
.result = "\xe2\x2e\x44\xdf\xd3\x60\x6d\xb2"
"\x70\x57\x37\xc5\xc2\x4f\x8d\x14"
"\xc6\xbf\x8b\xec\xf5\x62\x67\xf2"
"\x2f\xa1\xe6\xd6\xa7\xb1\x8c\x54"
"\xe5\x6b\x49\xf9\x6e\x90\xc3\xaa"
"\x7a\x00\x2e\x4d\x7f\x31\x2e\x81",
.rlen = 48,
}, {
.key = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76",
.klen = 16,
.iv = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
.assoc = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58"
"\x1a",
.alen = 33,
.input = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d\x4d\x54\x51\x84\x61\xf6\x8e"
"\x03\x31\xf2\x25\x16\xcc\xaa\xc6"
"\x75\x73\x20\x30\x59\x54\xb2\xf0"
"\x3a\x4b\xe0\x23\x8e\xa6\x08\x35"
"\x8a",
.ilen = 65,
.result = "\xc7\xca\x26\x61\x57\xee\xa2\xb9"
"\xb1\x37\xde\x95\x06\x90\x11\x08"
"\x4d\x30\x9f\x24\xc0\x56\xb7\xe1"
"\x0b\x9f\xd2\x57\xe9\xd2\xb1\x76"
"\x56\x9a\xb4\x58\xc5\x08\xfc\xb5"
"\xf2\x31\x9b\xc9\xcd\xb3\x64\xdb"
"\x6f\x50\xbf\xf4\x73\x9d\xfb\x6b"
"\xef\x35\x25\x48\xed\xcf\x29\xa8"
"\xac\xc3\xb9\xcb\x61\x8f\x73\x92"
"\x2c\x7a\x6f\xda\xf9\x09\x6f\xe1"
"\xc4",
.rlen = 81,
}, {
.key = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c",
.klen = 16,
.iv = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
.assoc = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e"
"\x28\xce\x57\x34\xcd\x6e\x84\x4c"
"\x17\x3c\xe1\xb2\xa8\x0b\xbb\xf1"
"\x96\x41\x0d\x69\xe8\x54\x0a\xc8"
"\x15\x4e\x91\x92\x89\x4b\xb7\x9b"
"\x21",
.alen = 65,
.input = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac",
.ilen = 33,
.result = "\x57\xcd\x3d\x46\xc5\xf9\x68\x3b"
"\x2c\x0f\xb4\x7e\x7b\x64\x3e\x40"
"\xf3\x78\x63\x34\x89\x79\x39\x6b"
"\x61\x64\x4a\x9a\xfa\x70\xa4\xd3"
"\x54\x0b\xea\x05\xa6\x95\x64\xed"
"\x3d\x69\xa2\x0c\x27\x56\x2f\x34"
"\x66",
.rlen = 49,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82",
.klen = 16,
.iv = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.assoc = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
.alen = 16,
.input = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.ilen = 16,
.result = "\xfc\x85\x06\x28\x8f\xe8\x23\x1f"
"\x33\x98\x87\xde\x08\xb6\xb6\xae"
"\x3e\xa4\xf8\x19\xf1\x92\x60\x39"
"\xb9\x6b\x3f\xdf\xc8\xcb\x30",
.rlen = 31,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
.klen = 16,
.iv = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
.assoc = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
.alen = 16,
.input = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.ilen = 16,
.result = "\x74\x7d\x70\x07\xe9\xba\x01\xee"
"\x6c\xc6\x6f\x50\x25\x33\xbe\x50"
"\x17\xb8\x17\x62\xed\x80\xa2\xf5"
"\x03\xde\x85\x71\x5d\x34",
.rlen = 30,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
.klen = 16,
.iv = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.assoc = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
.alen = 16,
.input = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.ilen = 16,
.result = "\xf4\xb3\x85\xf9\xac\xde\xb1\x38"
"\x29\xfd\x6c\x7c\x49\xe5\x1d\xaf"
"\xba\xea\xd4\xfa\x3f\x11\x33\x98",
.rlen = 24,
}, {
.key = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
.klen = 16,
.iv = "\xbb\x3a\xf7\x57\xc6\x36\x7c\x22"
"\x36\xab\xde\xc6\x6d\x32\x70\x17",
.assoc = "\xcb\x03\x85\xbf\x0a\xd5\x26\xa9"
"\x56\xe1\x0a\xeb\x6c\xfb\xa1\x98",
.alen = 16,
.input = "\xda\xcc\x14\x27\x4e\x74\xd1\x30"
"\x76\x18\x37\x0f\x6a\xc4\xd1\x1a",
.ilen = 16,
.result = "\xe6\x5c\x49\x4f\x78\xf3\x62\x86"
"\xe1\xb7\xa5\xc3\x32\x88\x3c\x8c"
"\x6e",
.rlen = 17,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec morus640_dec_tv_template[] = {
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.iv = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
.assoc = "",
.alen = 0,
.input = "\x89\x62\x7d\xf3\x07\x9d\x52\x05"
"\x53\xc3\x04\x60\x93\xb4\x37\x9a",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3c\x24\x39\x9f\x10\x7b\xa8\x1b"
"\x80\xda\xb2\x91\xf9\x24\xc2\x06",
.klen = 16,
.iv = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
.assoc = "",
.alen = 0,
.input = "\xa8\x8d\xe4\x90\xb5\x50\x8f\x78"
"\xb6\x10\x9a\x59\x5f\x61\x37\x70"
"\x09",
.ilen = 17,
.result = "\x69",
.rlen = 1,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x79\x49\x73\x3e\x20\xf7\x51\x37"
"\x01\xb4\x64\x22\xf3\x48\x85\x0c",
.klen = 16,
.iv = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
.assoc = "",
.alen = 0,
.input = "\x76\xdd\xb9\x05\x3d\xce\x61\x38"
"\xf3\xef\xf7\xe5\xd7\xfd\x70\xa5"
"\xcf\x9d\x64\xb8\x0a\x9f\xfd\x8b"
"\xd4\x6e\xfe\xd9\xc8\x63\x4b",
.ilen = 31,
.result = "\xa6\xa4\x1e\x76\xec\xd4\x50\xcc"
"\x62\x58\xe9\x8f\xef\xa4\x17",
.rlen = 15,
}, {
.key = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47\x12",
.klen = 16,
.iv = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
.assoc = "",
.alen = 0,
.input = "\xdc\x72\xe8\x14\xfb\x63\xad\x72"
"\x1f\x57\x9a\x1f\x88\x81\xdb\xd6"
"\xc1\x91\x9d\xb9\x25\xc4\x99\x4c"
"\x97\xcd\x8a\x0c\x9d\x68\x00\x1c",
.ilen = 32,
.result = "\xe2\xc9\x58\x15\xfc\x4f\xf8\xe8"
"\xe3\x32\x9b\x21\xe9\xc8\xd9\x97",
.rlen = 16,
}, {
.key = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.klen = 16,
.iv = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
.assoc = "",
.alen = 0,
.input = "\x6b\x4f\x3b\x90\x9a\xa2\xb3\x82"
"\x0a\xb8\x55\xee\xeb\x73\x4d\x7f"
"\x54\x11\x3a\x8a\x31\xa3\xb5\xf2"
"\xcd\x49\xdb\xf3\xee\x26\xbd\xa2"
"\x0d",
.ilen = 33,
.result = "\x1f\xee\x92\xb4\x0c\xcb\xa1\x04"
"\x64\x0c\x4d\xb2\xe3\xec\x9c\x9d"
"\x09",
.rlen = 17,
}, {
.key = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f",
.klen = 16,
.iv = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
.assoc = "",
.alen = 0,
.input = "\x59\xd1\x0f\x6b\xee\x27\x84\x92"
"\xb7\xa9\xb5\xdd\x02\xa4\x12\xa5"
"\x50\x32\xb4\x9a\x2e\x35\x83\x55"
"\x36\x12\x12\xed\xa3\x31\xc5\x30"
"\xa7\xe2\x4a\x6d\x05\x59\x43\x91"
"\x75\xfa\x6c\x17\xc6\x73\xca",
.ilen = 47,
.result = "\x5c\x13\xcb\x54\x1c\x47\x4a\x1f"
"\xe5\xe6\xff\x44\xdd\x11\x5f\xa3"
"\x33\xdd\xc2\xf8\xdd\x18\x2b\x93"
"\x57\x05\x01\x1c\x66\x22\xd3",
.rlen = 31,
}, {
.key = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25",
.klen = 16,
.iv = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
.assoc = "",
.alen = 0,
.input = "\xdb\x49\x68\x0f\x91\x5b\x21\xb1"
"\xcf\x50\xb2\x4c\x32\xe1\xa6\x69"
"\xc0\xfb\x44\x1f\xa0\x9a\xeb\x39"
"\x1b\xde\x68\x38\xcc\x27\x52\xc5"
"\xf6\x3e\x74\xea\x66\x5b\x5f\x0c"
"\x65\x9e\x58\xe6\x52\xa2\xfe\x59",
.ilen = 48,
.result = "\x98\x37\x05\xf3\x2c\xc2\xf3\x3b"
"\x66\xc0\xb1\xd5\xd7\x35\x21\xaa"
"\x5d\x9f\xce\x7c\xe2\xb8\xad\xad"
"\x19\x33\xe0\xf4\x40\x81\x72\x28",
.rlen = 32,
}, {
.key = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b",
.klen = 16,
.iv = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
.assoc = "\xc5",
.alen = 1,
.input = "\x56\xe7\x24\x52\xdd\x95\x60\x5b"
"\x09\x48\x39\x69\x9c\xb3\x62\x46",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\xe4\x25\xcd\xfa\x80\xdd\x46\xde"
"\x07\xd1\x90\x8b\xcf\x23\x15\x31",
.klen = 16,
.iv = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
.assoc = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76",
.alen = 15,
.input = "\xdd\xfa\x6c\x1f\x5d\x86\x87\x01"
"\x13\xe5\x73\x46\x46\xf2\x5c\xe1",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x20\x4a\x07\x99\x91\x58\xee\xfa"
"\x88\xab\x42\x1c\xc9\x47\xd7\x38",
.klen = 16,
.iv = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
.assoc = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b",
.alen = 16,
.input = "\xa6\x1b\xb9\xd7\x5e\x3c\xcf\xac"
"\xa9\x21\x45\x0b\x16\x52\xf7\xe1",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x5d\x6f\x41\x39\xa1\xd4\x97\x16"
"\x09\x85\xf4\xae\xc3\x6b\x9a\x3e",
.klen = 16,
.iv = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
.assoc = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c",
.alen = 17,
.input = "\x15\xff\xde\x3b\x34\xfc\xf6\xf9"
"\xbb\xa8\x62\xad\x0a\xf5\x48\x60",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x99\x93\x7a\xd8\xb1\x50\x40\x31"
"\x8a\x60\xa6\x3f\xbd\x90\x5d\x44",
.klen = 16,
.iv = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
.assoc = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
"\xaf\x0b\x02\x38\xcc\x44\xa7",
.alen = 31,
.input = "\xd2\x9d\xf8\x3b\xd7\x84\xe9\x2d"
"\x4b\xef\x75\x16\x0a\x99\xae\x6b",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\xd6\xb8\xb4\x77\xc1\xcb\xe9\x4d"
"\x0a\x3a\x58\xd1\xb7\xb4\x1f\x4a",
.klen = 16,
.iv = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
.assoc = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a",
.alen = 32,
.input = "\xe4\x8d\xa7\xa7\x45\xc1\x31\x4f"
"\xce\xfb\xaf\xd6\xc2\xe6\xee\xc0",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x12\xdd\xee\x17\xd1\x47\x92\x69"
"\x8b\x14\x0a\x62\xb1\xd9\xe2\x50",
.klen = 16,
.iv = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
.assoc = "\x31",
.alen = 1,
.input = "\xe2\x67\x38\x4f\xb9\xad\x7d\x38"
"\x01\xfe\x84\x14\x85\xf8\xd1\xe3"
"\x22",
.ilen = 17,
.result = "\x40",
.rlen = 1,
}, {
.key = "\x4f\x01\x27\xb6\xe1\xc3\x3a\x85"
"\x0c\xee\xbc\xf4\xab\xfd\xa5\x57",
.klen = 16,
.iv = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.assoc = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06",
.alen = 15,
.input = "\x77\x32\x61\xeb\xb4\x33\x29\x92"
"\x29\x95\xc5\x8e\x85\x76\xab\xfc"
"\x07\x95\xa7\x44\x74\xf7\x22\xff"
"\xd8\xd8\x36\x3d\x8a\x7f\x9e",
.ilen = 31,
.result = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37",
.rlen = 15,
}, {
.key = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67\x5d",
.klen = 16,
.iv = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
.assoc = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60",
.alen = 16,
.input = "\xd8\xfd\x44\x45\xf6\x42\x12\x38"
"\xf2\x0b\xea\x4f\x9e\x11\x61\x07"
"\x48\x67\x98\x18\x9b\xd0\x0c\x59"
"\x67\xa4\x11\xb3\x2b\xd6\xc1\x70",
.ilen = 32,
.result = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2",
.rlen = 16,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.klen = 16,
.iv = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
.assoc = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b",
.alen = 17,
.input = "\xb1\xab\x53\x4e\xc7\x40\x16\xb6"
"\x71\x3a\x00\x9f\x41\x88\xb0\xb2"
"\x71\x83\x85\x5f\xc8\x79\x0a\x99"
"\x99\xdc\x89\x1c\x88\xd2\x3e\xf9"
"\x83",
.ilen = 33,
.result = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05",
.rlen = 17,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69",
.klen = 16,
.iv = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
.assoc = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
"\x38\x1d\x3b\x4a\xe9\x7e\x62",
.alen = 31,
.input = "\x29\xc4\xf0\x03\xc1\x86\xdf\x06"
"\x5c\x7b\xef\x64\x87\x00\xd1\x37"
"\xa7\x08\xbc\x7f\x8f\x41\x54\xd0"
"\x3e\xf1\xc3\xa2\x96\x84\xdd\x2a"
"\x2d\x21\x30\xf9\x02\xdb\x06\x0c"
"\xf1\x5a\x66\x69\xe0\xca\x83",
.ilen = 47,
.result = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a",
.rlen = 31,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70",
.klen = 16,
.iv = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
.assoc = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81",
.alen = 32,
.input = "\xe2\x2e\x44\xdf\xd3\x60\x6d\xb2"
"\x70\x57\x37\xc5\xc2\x4f\x8d\x14"
"\xc6\xbf\x8b\xec\xf5\x62\x67\xf2"
"\x2f\xa1\xe6\xd6\xa7\xb1\x8c\x54"
"\xe5\x6b\x49\xf9\x6e\x90\xc3\xaa"
"\x7a\x00\x2e\x4d\x7f\x31\x2e\x81",
.ilen = 48,
.result = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37",
.rlen = 32,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76",
.klen = 16,
.iv = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
.assoc = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58"
"\x1a",
.alen = 33,
.input = "\xc7\xca\x26\x61\x57\xee\xa2\xb9"
"\xb1\x37\xde\x95\x06\x90\x11\x08"
"\x4d\x30\x9f\x24\xc0\x56\xb7\xe1"
"\x0b\x9f\xd2\x57\xe9\xd2\xb1\x76"
"\x56\x9a\xb4\x58\xc5\x08\xfc\xb5"
"\xf2\x31\x9b\xc9\xcd\xb3\x64\xdb"
"\x6f\x50\xbf\xf4\x73\x9d\xfb\x6b"
"\xef\x35\x25\x48\xed\xcf\x29\xa8"
"\xac\xc3\xb9\xcb\x61\x8f\x73\x92"
"\x2c\x7a\x6f\xda\xf9\x09\x6f\xe1"
"\xc4",
.ilen = 81,
.result = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d\x4d\x54\x51\x84\x61\xf6\x8e"
"\x03\x31\xf2\x25\x16\xcc\xaa\xc6"
"\x75\x73\x20\x30\x59\x54\xb2\xf0"
"\x3a\x4b\xe0\x23\x8e\xa6\x08\x35"
"\x8a",
.rlen = 65,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c",
.klen = 16,
.iv = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
.assoc = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e"
"\x28\xce\x57\x34\xcd\x6e\x84\x4c"
"\x17\x3c\xe1\xb2\xa8\x0b\xbb\xf1"
"\x96\x41\x0d\x69\xe8\x54\x0a\xc8"
"\x15\x4e\x91\x92\x89\x4b\xb7\x9b"
"\x21",
.alen = 65,
.input = "\x57\xcd\x3d\x46\xc5\xf9\x68\x3b"
"\x2c\x0f\xb4\x7e\x7b\x64\x3e\x40"
"\xf3\x78\x63\x34\x89\x79\x39\x6b"
"\x61\x64\x4a\x9a\xfa\x70\xa4\xd3"
"\x54\x0b\xea\x05\xa6\x95\x64\xed"
"\x3d\x69\xa2\x0c\x27\x56\x2f\x34"
"\x66",
.ilen = 49,
.result = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac",
.rlen = 33,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82",
.klen = 16,
.iv = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.assoc = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85",
.alen = 16,
.input = "\xfc\x85\x06\x28\x8f\xe8\x23\x1f"
"\x33\x98\x87\xde\x08\xb6\xb6\xae"
"\x3e\xa4\xf8\x19\xf1\x92\x60\x39"
"\xb9\x6b\x3f\xdf\xc8\xcb\x30",
.ilen = 31,
.result = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07",
.rlen = 16,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
.klen = 16,
.iv = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
.assoc = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c",
.alen = 16,
.input = "\x74\x7d\x70\x07\xe9\xba\x01\xee"
"\x6c\xc6\x6f\x50\x25\x33\xbe\x50"
"\x17\xb8\x17\x62\xed\x80\xa2\xf5"
"\x03\xde\x85\x71\x5d\x34",
.ilen = 30,
.result = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d",
.rlen = 16,
}, {
.key = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
.klen = 16,
.iv = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.assoc = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92",
.alen = 16,
.input = "\xf4\xb3\x85\xf9\xac\xde\xb1\x38"
"\x29\xfd\x6c\x7c\x49\xe5\x1d\xaf"
"\xba\xea\xd4\xfa\x3f\x11\x33\x98",
.ilen = 24,
.result = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13",
.rlen = 16,
}, {
.key = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
.klen = 16,
.iv = "\xbb\x3a\xf7\x57\xc6\x36\x7c\x22"
"\x36\xab\xde\xc6\x6d\x32\x70\x17",
.assoc = "\xcb\x03\x85\xbf\x0a\xd5\x26\xa9"
"\x56\xe1\x0a\xeb\x6c\xfb\xa1\x98",
.alen = 16,
.input = "\xe6\x5c\x49\x4f\x78\xf3\x62\x86"
"\xe1\xb7\xa5\xc3\x32\x88\x3c\x8c"
"\x6e",
.ilen = 17,
.result = "\xda\xcc\x14\x27\x4e\x74\xd1\x30"
"\x76\x18\x37\x0f\x6a\xc4\xd1\x1a",
.rlen = 16,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* MORUS-1280 test vectors - generated via reference implementation from
* SUPERCOP (https://bench.cr.yp.to/supercop.html):
*
* https://bench.cr.yp.to/supercop/supercop-20170228.tar.xz
* (see crypto_aead/morus1280128v2/ and crypto_aead/morus1280256v2/ )
*/
static const struct aead_testvec morus1280_enc_tv_template[] = {
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.iv = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
.assoc = "",
.alen = 0,
.input = "",
.ilen = 0,
.result = "\x91\x85\x0f\xf5\x52\x9e\xce\xce"
"\x65\x99\xc7\xbf\xd3\x76\xe8\x98",
.rlen = 16,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3c\x24\x39\x9f\x10\x7b\xa8\x1b"
"\x80\xda\xb2\x91\xf9\x24\xc2\x06",
.klen = 16,
.iv = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
.assoc = "",
.alen = 0,
.input = "\x69",
.ilen = 1,
.result = "\x88\xc3\x4c\xf0\x2f\x43\x76\x13"
"\x96\xda\x76\x34\x33\x4e\xd5\x39"
"\x73",
.rlen = 17,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x79\x49\x73\x3e\x20\xf7\x51\x37"
"\x01\xb4\x64\x22\xf3\x48\x85\x0c",
.klen = 16,
.iv = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
.assoc = "",
.alen = 0,
.input = "\xa6\xa4\x1e\x76\xec\xd4\x50\xcc"
"\x62\x58\xe9\x8f\xef\xa4\x17\x91"
"\xb4\x96\x9f\x6b\xce\x38\xa5\x46"
"\x13\x7d\x64\x93\xd7\x05\xf5",
.ilen = 31,
.result = "\x3e\x5c\x3b\x58\x3b\x7d\x2a\x22"
"\x75\x0b\x24\xa6\x0e\xc3\xde\x52"
"\x97\x0b\x64\xd4\xce\x90\x52\xf7"
"\xef\xdb\x6a\x38\xd2\xa8\xa1\x0d"
"\xe0\x61\x33\x24\xc6\x4d\x51\xbc"
"\xa4\x21\x74\xcf\x19\x16\x59",
.rlen = 47,
crypto: drbg - DRBG testmgr test vectors All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors. In addition, all permutations of use cases of the DRBG are covered: * with and without predition resistance * with and without additional information string * with and without personalization string As the DRBG implementation is agnositc of the specific backend cipher, only test vectors for one specific backend cipher is used. For example: the Hash DRBG uses the same code paths irrespectively of using SHA-256 or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all DRBG code paths of SHA-512. Signed-off-by: Stephan Mueller <smueller@chronox.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-05-31 09:24:38 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47\x12",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
.assoc = "",
.alen = 0,
.input = "\xe2\xc9\x58\x15\xfc\x4f\xf8\xe8"
"\xe3\x32\x9b\x21\xe9\xc8\xd9\x97"
"\xde\x58\xab\xf0\xd3\xd8\x27\x60"
"\xd5\xaa\x43\x6b\xb1\x64\x95\xa4",
.ilen = 32,
.result = "\x30\x82\x9c\x2b\x67\xcb\xf9\x1f"
"\xde\x9f\x77\xb2\xda\x92\x61\x5c"
"\x09\x0b\x2d\x9a\x26\xaa\x1c\x06"
"\xab\x74\xb7\x2b\x95\x5f\x9f\xa1"
"\x9a\xff\x50\xa0\xa2\xff\xc5\xad"
"\x21\x8e\x84\x5c\x12\x61\xb2\xae",
.rlen = 48,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.klen = 16,
.iv = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
.assoc = "",
.alen = 0,
.input = "\x1f\xee\x92\xb4\x0c\xcb\xa1\x04"
"\x64\x0c\x4d\xb2\xe3\xec\x9c\x9d"
"\x09\x1a\xb7\x74\xd8\x78\xa9\x79"
"\x96\xd8\x22\x43\x8c\xc3\x34\x7b"
"\xc4",
.ilen = 33,
.result = "\x67\x5d\x8e\x45\xc8\x39\xf5\x17"
"\xc1\x1d\x2a\xdd\x88\x67\xda\x1f"
"\x6d\xe8\x37\x28\x5a\xc1\x5e\x9f"
"\xa6\xec\xc6\x92\x05\x4b\xc0\xa3"
"\x63\xef\x88\xa4\x9b\x0a\x5c\xed"
"\x2b\x6a\xac\x63\x52\xaa\x10\x94"
"\xd0",
.rlen = 49,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f",
crypto: testmgr - add larger cast5 testvectors New ECB, CBC and CTR testvectors for cast5. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:37:21 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
.assoc = "",
.alen = 0,
.input = "\x5c\x13\xcb\x54\x1c\x47\x4a\x1f"
"\xe5\xe6\xff\x44\xdd\x11\x5f\xa3"
"\x33\xdd\xc2\xf8\xdd\x18\x2b\x93"
"\x57\x05\x01\x1c\x66\x22\xd3\x51"
"\xd3\xdf\x18\xc9\x30\x66\xed\xb1"
"\x96\x58\xd5\x8c\x64\x8c\x7c\xf5"
"\x01\xd0\x74\x5f\x9b\xaa\xf6\xd1"
"\xe6\x16\xa2\xac\xde\x47\x40",
.ilen = 63,
.result = "\x7d\x61\x1a\x35\x20\xcc\x07\x88"
"\x03\x98\x87\xcf\xc0\x6e\x4d\x19"
"\xe3\xd4\x0b\xfb\x29\x8f\x49\x1a"
"\x3a\x06\x77\xce\x71\x2c\xcd\xdd"
"\xed\xf6\xc9\xbe\xa6\x3b\xb8\xfc"
"\x6c\xbe\x77\xed\x74\x0e\x20\x85"
"\xd0\x65\xde\x24\x6f\xe3\x25\xc5"
"\xdf\x5b\x0f\xbd\x8a\x88\x78\xc9"
"\xe5\x81\x37\xde\x84\x7a\xf6\x84"
"\x99\x7a\x72\x9c\x54\x31\xa1",
.rlen = 79,
}, {
.key = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
.assoc = "",
.alen = 0,
.input = "\x98\x37\x05\xf3\x2c\xc2\xf3\x3b"
"\x66\xc0\xb1\xd5\xd7\x35\x21\xaa"
"\x5d\x9f\xce\x7c\xe2\xb8\xad\xad"
"\x19\x33\xe0\xf4\x40\x81\x72\x28"
"\xe1\x8b\x1c\xf8\x91\x78\xff\xaf"
"\xb0\x68\x69\xf2\x27\x35\x91\x84"
"\x2e\x37\x5b\x00\x04\xff\x16\x9c"
"\xb5\x19\x39\xeb\xd9\xcd\x29\x9a",
.ilen = 64,
.result = "\x05\xc5\xb1\xf9\x1b\xb9\xab\x2c"
"\xa5\x07\x12\xa7\x12\x39\x60\x66"
"\x30\x81\x4a\x03\x78\x28\x45\x52"
"\xd2\x2b\x24\xfd\x8b\xa5\xb7\x66"
"\x6f\x45\xd7\x3b\x67\x6f\x51\xb9"
"\xc0\x3d\x6c\xca\x1e\xae\xff\xb6"
"\x79\xa9\xe4\x82\x5d\x4c\x2d\xdf"
"\xeb\x71\x40\xc9\x2c\x40\x45\x6d"
"\x73\x77\x01\xf3\x4f\xf3\x9d\x2a"
"\x5d\x57\xa8\xa1\x18\xa2\xad\xcb",
.rlen = 80,
}, {
.key = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b",
.klen = 16,
.iv = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
.assoc = "\xc5",
.alen = 1,
.input = "",
.ilen = 0,
.result = "\x4d\xbf\x11\xac\x7f\x97\x0b\x2e"
"\x89\x3b\x9d\x0f\x83\x1c\x08\xc3",
.rlen = 16,
}, {
.key = "\xe4\x25\xcd\xfa\x80\xdd\x46\xde"
"\x07\xd1\x90\x8b\xcf\x23\x15\x31",
.klen = 16,
.iv = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
.assoc = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34"
"\xe8\x73\x62\x64\xab\x50\xd0\xda"
"\x6b\x83\x66\xaf\x3e\x27\xc9",
.alen = 31,
.input = "",
.ilen = 0,
.result = "\x5b\xc0\x8d\x54\xe4\xec\xbe\x38"
"\x03\x12\xf9\xcc\x9e\x46\x42\x92",
.rlen = 16,
}, {
.key = "\x20\x4a\x07\x99\x91\x58\xee\xfa"
"\x88\xab\x42\x1c\xc9\x47\xd7\x38",
.klen = 16,
.iv = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
.assoc = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b"
"\x12\x35\x6e\xe8\xb0\xf0\x52\xf3"
"\x2d\xb0\x45\x87\x18\x86\x68\xf6",
.alen = 32,
.input = "",
.ilen = 0,
.result = "\x48\xc5\xc3\x4c\x40\x2e\x2f\xc2"
"\x6d\x65\xe0\x67\x9c\x1d\xa0\xf0",
.rlen = 16,
}, {
.key = "\x5d\x6f\x41\x39\xa1\xd4\x97\x16"
"\x09\x85\xf4\xae\xc3\x6b\x9a\x3e",
.klen = 16,
.iv = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
.assoc = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c\xf8\x79\x6c\xb6\x90\xd4\x0d"
"\xee\xde\x23\x60\xf2\xe5\x08\xcc"
"\x97",
.alen = 33,
.input = "",
.ilen = 0,
.result = "\x28\x64\x78\x51\x55\xd8\x56\x4a"
"\x58\x3e\xf7\xbe\xee\x21\xfe\x94",
.rlen = 16,
}, {
.key = "\x99\x93\x7a\xd8\xb1\x50\x40\x31"
"\x8a\x60\xa6\x3f\xbd\x90\x5d\x44",
.klen = 16,
.iv = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
.assoc = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
"\xaf\x0b\x02\x38\xcc\x44\xa7\xa3"
"\xa6\xbf\x31\x93\x60\xcd\xda\x63"
"\x2c\xb1\xaa\x19\xc8\x19\xf8\xeb"
"\x03\xa1\xe8\xbe\x37\x54\xec\xa2"
"\xcd\x2c\x45\x58\xbd\x8e\x80",
.alen = 63,
.input = "",
.ilen = 0,
.result = "\xb3\xa6\x00\x4e\x09\x20\xac\x21"
"\x77\x72\x69\x76\x2d\x36\xe5\xc8",
.rlen = 16,
}, {
.key = "\xd6\xb8\xb4\x77\xc1\xcb\xe9\x4d"
"\x0a\x3a\x58\xd1\xb7\xb4\x1f\x4a",
.klen = 16,
.iv = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
.assoc = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a"
"\xb4\x6b\x35\xc2\xc1\xdf\xed\x60"
"\x46\xc1\x3e\x7f\x8c\xc2\x0e\x7a"
"\x30\x08\xd0\x5f\xa0\xaa\x0c\x6d"
"\x9c\x2f\xdb\x97\xb8\x15\x69\x01",
.alen = 64,
.input = "",
.ilen = 0,
.result = "\x65\x33\x7b\xa1\x63\xf4\x20\xdd"
"\xe4\xb9\x4a\xaa\x9a\x21\xaa\x14",
.rlen = 16,
}, {
.key = "\x12\xdd\xee\x17\xd1\x47\x92\x69"
"\x8b\x14\x0a\x62\xb1\xd9\xe2\x50",
.klen = 16,
.iv = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
.assoc = "\x31",
.alen = 1,
.input = "\x40",
.ilen = 1,
.result = "\x1d\x47\x17\x34\x86\xf5\x54\x1a"
"\x6d\x28\xb8\x5d\x6c\xcf\xa0\xb9"
"\xbf",
.rlen = 17,
}, {
.key = "\x4f\x01\x27\xb6\xe1\xc3\x3a\x85"
"\x0c\xee\xbc\xf4\xab\xfd\xa5\x57",
.klen = 16,
.iv = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.assoc = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a"
"\xe6\x01\xa8\x7e\xca\x10\xdc\x73"
"\xf4\x94\x9f\xc1\x5a\x61\x85",
.alen = 31,
.input = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37\xdb"
"\xb0\xb2\x2b\x9f\x0b\xb8\xbd\x7a"
"\x24\xa0\xd6\xb7\x11\x79\x6c",
.ilen = 31,
.result = "\x78\x90\x52\xae\x0f\xf7\x2e\xef"
"\x63\x09\x08\x58\xb5\x56\xbd\x72"
"\x6e\x42\xcf\x27\x04\x7c\xdb\x92"
"\x18\xe9\xa4\x33\x90\xba\x62\xb5"
"\x70\xd3\x88\x9b\x4f\x05\xa7\x51"
"\x85\x87\x17\x09\x42\xed\x4e",
.rlen = 47,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67\x5d",
.klen = 16,
.iv = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
.assoc = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60"
"\x10\xc3\xb3\x02\xcf\xb0\x5e\x8d"
"\xb5\xc2\x7e\x9a\x35\xc0\x24\xfd",
.alen = 32,
.input = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2"
"\xdb\x74\x36\x23\x11\x58\x3f\x93"
"\xe5\xcd\xb5\x90\xeb\xd8\x0c\xb3",
.ilen = 32,
.result = "\x1d\x2c\x57\xe0\x50\x38\x3d\x41"
"\x2e\x71\xc8\x3b\x92\x43\x58\xaf"
"\x5a\xfb\xad\x8f\xd9\xd5\x8a\x5e"
"\xdb\xf3\xcd\x3a\x2b\xe1\x2c\x1a"
"\xb0\xed\xe3\x0c\x6e\xf9\xf2\xd6"
"\x90\xe6\xb1\x0e\xa5\x8a\xac\xb7",
.rlen = 48,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
crypto: testmgr - add larger cast5 testvectors New ECB, CBC and CTR testvectors for cast5. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:37:21 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
.assoc = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b\x86\xbf\x86\xd4\x50\xe0\xa7"
"\x76\xef\x5c\x72\x0f\x1f\xc3\xd4"
"\xee",
.alen = 33,
.input = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05\x36\x42\xa7\x16\xf8\xc1\xad"
"\xa7\xfb\x94\x68\xc5\x37\xab\x8a"
"\x72",
.ilen = 33,
.result = "\x59\x10\x84\x1c\x83\x4c\x8b\xfc"
"\xfd\x2e\x4b\x46\x84\xff\x78\x4e"
"\x50\xda\x5c\xb9\x61\x1d\xf5\xb9"
"\xfe\xbb\x7f\xae\x8c\xc1\x24\xbd"
"\x8c\x6f\x1f\x9b\xce\xc6\xc1\x37"
"\x08\x06\x5a\xe5\x96\x10\x95\xc2"
"\x5e",
.rlen = 49,
}, {
.key = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69",
crypto: testmgr - add larger cast5 testvectors New ECB, CBC and CTR testvectors for cast5. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:37:21 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
.assoc = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
"\x38\x1d\x3b\x4a\xe9\x7e\x62\xaa"
"\xfd\xc9\x4a\xa9\xa9\x39\x4b\x54"
"\xc8\x0e\x24\x7f\x5e\x10\x7a\x45"
"\x10\x0b\x56\x85\xad\x54\xaa\x66"
"\xa8\x43\xcd\xd4\x9b\xb7\xfa",
.alen = 63,
.input = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a\x60"
"\x80\xf4\x4b\xf4\xc1\x3d\xd0\x93"
"\xcf\x12\xc9\x59\x8f\x7a\x7f\xa8"
"\x1b\xa5\x50\xed\x87\xa9\x72\x59"
"\x9c\x44\xb2\xa4\x99\x98\x34",
.ilen = 63,
.result = "\x9a\x12\xbc\xdf\x72\xa8\x56\x22"
"\x49\x2d\x07\x92\xfc\x3d\x6d\x5f"
"\xef\x36\x19\xae\x91\xfa\xd6\x63"
"\x46\xea\x8a\x39\x14\x21\xa6\x37"
"\x18\xfc\x97\x3e\x16\xa5\x4d\x39"
"\x45\x2e\x69\xcc\x9c\x5f\xdf\x6d"
"\x5e\xa2\xbf\xac\x83\x32\x72\x52"
"\x58\x58\x23\x40\xfd\xa5\xc2\xe6"
"\xe9\x5a\x50\x98\x00\x58\xc9\x86"
"\x4f\x20\x37\xdb\x7b\x22\xa3",
.rlen = 79,
}, {
.key = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70",
crypto: testmgr - add larger cast5 testvectors New ECB, CBC and CTR testvectors for cast5. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:37:21 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
.assoc = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81"
"\x0b\x76\x4f\xd7\x0a\x4b\x5e\x51"
"\xe3\x1d\xb9\xe5\x21\xb9\x8f\xd4"
"\x3d\x72\x3e\x26\x16\xa9\xca\x32"
"\x77\x47\x63\x14\x95\x3d\xe4\x34",
.alen = 64,
.input = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37"
"\x8f\xa1\x50\x23\x22\x4f\xe3\x91"
"\xe9\x21\x5e\xbf\x52\x23\x95\x37"
"\x48\x0c\x38\x8f\xf0\xff\x92\x24"
"\x6b\x47\x49\xe3\x94\x1f\x1e\x01",
.ilen = 64,
.result = "\xe6\xeb\x92\x5a\x5b\xf0\x2d\xbb"
"\x23\xec\x35\xe3\xae\xc9\xfb\x0b"
"\x90\x14\x46\xeb\xa8\x8d\xb0\x9b"
"\x39\xda\x8b\x48\xec\xb2\x00\x4e"
"\x80\x6f\x46\x4f\x9b\x1e\xbb\x35"
"\xea\x5a\xbc\xa2\x36\xa5\x89\x45"
"\xc2\xd6\xd7\x15\x0b\xf6\x6c\x56"
"\xec\x99\x7d\x61\xb3\x15\x93\xed"
"\x83\x1e\xd9\x48\x84\x0b\x37\xfe"
"\x95\x74\x44\xd5\x54\xa6\x27\x06",
.rlen = 80,
}, {
.key = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76",
crypto: testmgr - add test vectors for partial ctr(cast5) and ctr(cast6) Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:54 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
.assoc = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58"
"\x1a\x22\x53\x05\x6b\x5c\x71\x4f"
"\xfd\x2d\x4d\x4c\xe5\x62\xa5\x63"
"\x6a\xda\x26\xc8\x7f\xff\xea\xfd"
"\x46\x4a\xfa\x53\x8f\xc4\xcd\x68"
"\x58",
.alen = 65,
.input = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d\x4d\x54\x51\x84\x61\xf6\x8e"
"\x03\x31\xf2\x25\x16\xcc\xaa\xc6"
"\x75\x73\x20\x30\x59\x54\xb2\xf0"
"\x3a\x4b\xe0\x23\x8e\xa6\x08\x35"
"\x8a\xdf\x27\xa0\xe4\x60\x99\xae"
"\x8e\x43\xd9\x39\x7b\x10\x40\x67"
"\x5c\x7e\xc9\x70\x63\x34\xca\x59"
"\xfe\x86\xbc\xb7\x9c\x39\xf3\x6d"
"\x6a\x41\x64\x6f\x16\x7f\x65\x7e"
"\x89\x84\x68\xeb\xb0\x51\xbe\x55"
"\x33\x16\x59\x6c\x3b\xef\x88\xad"
"\x2f\xab\xbc\x25\x76\x87\x41\x2f"
"\x36",
.ilen = 129,
.result = "\x89\x24\x27\x86\xdc\xd7\x6b\xd9"
"\xd1\xcd\xdc\x16\xdd\x2c\xc1\xfb"
"\x52\xb5\xb3\xab\x50\x99\x3f\xa0"
"\x38\xa4\x74\xa5\x04\x15\x63\x05"
"\x8f\x54\x81\x06\x5a\x6b\xa4\x63"
"\x6d\xa7\x21\xcb\xff\x42\x30\x8e"
"\x3b\xd1\xca\x3f\x4b\x1a\xb8\xc3"
"\x42\x01\xe6\xbc\x75\x15\x87\xee"
"\xc9\x8e\x65\x01\xd9\xd8\xb5\x9f"
"\x48\x86\xa6\x5f\x2c\xc7\xb5\xb0"
"\xed\x5d\x14\x7c\x3f\x40\xb1\x0b"
"\x72\xef\x94\x8d\x7a\x85\x56\xe5"
"\x56\x08\x15\x56\xba\xaf\xbd\xf0"
"\x20\xef\xa0\xf6\xa9\xad\xa2\xc9"
"\x1c\x3b\x28\x51\x7e\x77\xb2\x18"
"\x4f\x61\x64\x37\x22\x36\x6d\x78"
"\xed\xed\x35\xe8\x83\xa5\xec\x25"
"\x6b\xff\x5f\x1a\x09\x96\x3d\xdc"
"\x20",
.rlen = 145,
}, {
.key = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c",
crypto: testmgr - add larger cast5 testvectors New ECB, CBC and CTR testvectors for cast5. We need larger testvectors to check parallel code paths in the optimized implementation. Tests have also been added to the tcrypt module. Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-07-11 11:37:21 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
.assoc = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e"
"\x28\xce\x57\x34\xcd\x6e\x84\x4c"
"\x17\x3c\xe1\xb2\xa8\x0b\xbb\xf1"
"\x96\x41\x0d\x69\xe8\x54\x0a\xc8"
"\x15\x4e\x91\x92\x89\x4b\xb7\x9b"
"\x21\xf7\x42\x89\xac\x12\x2a\x54"
"\x69\xee\x18\xc7\x8d\xed\xe8\xfd"
"\xbb\x04\x28\xe6\x8a\x3c\x98\xc1"
"\x04\x2d\xa9\xa1\x24\x83\xff\xe9"
"\x55\x7a\xf0\xd1\xf6\x63\x05\xe1"
"\xd9\x1e\x75\x72\xc1\x9f\xae\x32"
"\xe1\x6b\xcd\x9e\x61\x19\x23\x86"
"\xd9\xd2\xaf\x8e\xd5\xd3\xa8\xa9"
"\x51",
.alen = 129,
.input = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
.ilen = 65,
.result = "\x36\x78\xb9\x22\xde\x62\x35\x55"
"\x1a\x7a\xf5\x45\xbc\xd7\x15\x82"
"\x01\xe9\x5a\x07\xea\x46\xaf\x91"
"\xcb\x73\xa5\xee\xe1\xb4\xbf\xc2"
"\xdb\xd2\x9d\x59\xde\xfc\x83\x00"
"\xf5\x46\xac\x97\xd5\x57\xa9\xb9"
"\x1f\x8c\xe8\xca\x68\x8b\x91\x0c"
"\x01\xbe\x0a\xaf\x7c\xf6\x67\xa4"
"\xbf\xbc\x88\x3f\x5d\xd1\xf9\x19"
"\x0f\x9d\xb2\xaf\xb9\x6e\x17\xdf"
"\xa2",
.rlen = 81,
}, {
.key = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82",
crypto: testmgr - add test vectors for partial ctr(cast5) and ctr(cast6) Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:54 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.assoc = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85"
"\x0e\x51\xf9\x1c\xee\x70\x6a\x27"
"\x3d\xd3\xb7\xac\x51\xfa\xdf\x05",
.alen = 32,
.input = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07"
"\xd9\x02\x7c\x3d\x2f\x18\x4b\x2d"
"\x6e\xde\xee\xa2\x08\x12\xc7\xba",
.ilen = 32,
.result = "\x08\x1b\x95\x0e\x41\x95\x02\x4b"
"\x9c\xbb\xa8\xd0\x7c\xd3\x44\x6e"
"\x89\x14\x33\x70\x0a\xbc\xea\x39"
"\x88\xaa\x2b\xd5\x73\x11\x55\xf5"
"\x33\x33\x9c\xd7\x42\x34\x49\x8e"
"\x2f\x03\x30\x05\x47\xaf\x34",
.rlen = 47,
}, {
.key = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
.klen = 16,
.iv = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
.assoc = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c"
"\x39\x14\x05\xa0\xf3\x10\xec\x41"
"\xff\x01\x95\x84\x2b\x59\x7f\xdb",
.alen = 32,
.input = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d"
"\x03\xc4\x88\xc1\x35\xb8\xcd\x47"
"\x2f\x0c\xcd\x7a\xe2\x71\x66\x91",
.ilen = 32,
.result = "\x97\xca\xf4\xe0\x8d\x89\xbf\x68"
"\x0c\x60\xb9\x27\xdf\xaa\x41\xc6"
"\x25\xd8\xf7\x1f\x10\x15\x48\x61"
"\x4c\x95\x00\xdf\x51\x9b\x7f\xe6"
"\x24\x40\x9e\xbe\x3b\xeb\x1b\x98"
"\xb9\x9c\xe5\xef\xf2\x05",
.rlen = 46,
}, {
.key = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
.klen = 16,
.iv = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.assoc = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92"
"\x63\xd6\x10\x24\xf8\xb0\x6e\x5a"
"\xc0\x2e\x74\x5d\x06\xb8\x1e\xb2",
.alen = 32,
.input = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13"
"\x2e\x86\x93\x45\x3a\x58\x4f\x61"
"\xf0\x3a\xac\x53\xbc\xd0\x06\x68",
.ilen = 32,
.result = "\x63\x4c\x2a\x8e\xb4\x6b\x63\x0d"
"\xb5\xec\x9b\x4e\x12\x23\xa3\xcf"
"\x1a\x5a\x70\x15\x5a\x10\x40\x51"
"\xca\x47\x4c\x9d\xc9\x97\xf4\x77"
"\xdb\xc8\x10\x2d\xdc\x65\x20\x3f",
.rlen = 40,
}, {
.key = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
.klen = 16,
.iv = "\xbb\x3a\xf7\x57\xc6\x36\x7c\x22"
"\x36\xab\xde\xc6\x6d\x32\x70\x17",
.assoc = "\xcb\x03\x85\xbf\x0a\xd5\x26\xa9"
"\x56\xe1\x0a\xeb\x6c\xfb\xa1\x98"
"\x8d\x98\x1c\xa8\xfe\x50\xf0\x74"
"\x81\x5c\x53\x35\xe0\x17\xbd\x88",
.alen = 32,
.input = "\xda\xcc\x14\x27\x4e\x74\xd1\x30"
"\x76\x18\x37\x0f\x6a\xc4\xd1\x1a"
"\x58\x49\x9f\xc9\x3f\xf8\xd1\x7a"
"\xb2\x67\x8b\x2b\x96\x2f\xa5\x3e",
.ilen = 32,
.result = "\xf1\x62\x44\xc7\x5f\x19\xca\x43"
"\x47\x2c\xaf\x68\x82\xbd\x51\xef"
"\x3d\x65\xd8\x45\x2d\x06\x07\x78"
"\x08\x2e\xb3\x23\xcd\x81\x12\x55"
"\x1a",
.rlen = 33,
}, {
.key = "\xe9\x95\xa2\x8f\x93\x13\x7b\xb7"
"\x96\x4e\x63\x33\x69\x8d\x02\x9b"
"\x23\xf9\x22\xeb\x80\xa0\xb1\x81"
"\xe2\x73\xc3\x21\x4d\x47\x8d\xf4",
.klen = 32,
.iv = "\xf8\x5e\x31\xf7\xd7\xb2\x25\x3e"
"\xb7\x85\x90\x58\x67\x57\x33\x1d",
.assoc = "",
.alen = 0,
.input = "",
.ilen = 0,
.result = "\xdf\x2f\x83\xc0\x45\x4a\x2c\xcf"
"\xb9\xd2\x41\xf6\x80\xa1\x52\x70",
.rlen = 16,
}, {
.key = "\x25\xba\xdc\x2e\xa3\x8f\x24\xd3"
"\x17\x29\x15\xc5\x63\xb2\xc5\xa1"
"\x4d\xbc\x2d\x6f\x85\x40\x33\x9a"
"\xa3\xa0\xa1\xfa\x27\xa6\x2c\xca",
.klen = 32,
.iv = "\x34\x83\x6a\x96\xe7\x2d\xce\x5a"
"\x38\x5f\x42\xe9\x61\x7b\xf5\x23",
.assoc = "",
.alen = 0,
.input = "\x53",
.ilen = 1,
.result = "\x01\xd8\x55\x3c\xc0\x5a\x4b\xc7"
"\x01\xf4\x08\xe3\x0d\xf7\xf0\x78"
"\x53",
.rlen = 17,
}, {
.key = "\x62\xdf\x16\xcd\xb3\x0a\xcc\xef"
"\x98\x03\xc7\x56\x5d\xd6\x87\xa8"
"\x77\x7e\x39\xf3\x8a\xe0\xb5\xb4"
"\x65\xce\x80\xd2\x01\x05\xcb\xa1",
.klen = 32,
.iv = "\x71\xa8\xa4\x35\xf7\xa9\x76\x75"
"\xb8\x39\xf4\x7a\x5b\x9f\xb8\x29",
.assoc = "",
.alen = 0,
.input = "\x8f\x3a\xc1\x05\x7f\xe7\xcb\x83"
"\xf9\xa6\x4d\xc3\x58\x31\x19\x2c"
"\xd7\x90\xc2\x56\x4e\xd8\x57\xc7"
"\xf6\xf0\x27\xb4\x25\x4c\x83",
.ilen = 31,
.result = "\xc2\x4b\x41\x0f\x2d\xb9\x62\x07"
"\xff\x8e\x74\xf8\xa1\xa6\xd5\x37"
"\xa5\x64\x31\x5c\xca\x73\x9b\x43"
"\xe6\x70\x63\x46\x95\xcb\xf7\xb5"
"\x20\x8c\x75\x7a\x2a\x17\x2f\xa9"
"\xb8\x4d\x11\x42\xd1\xf8\xf1",
.rlen = 47,
}, {
.key = "\x9e\x03\x4f\x6d\xc3\x86\x75\x0a"
"\x19\xdd\x79\xe8\x57\xfb\x4a\xae"
"\xa2\x40\x45\x77\x90\x80\x37\xce"
"\x26\xfb\x5f\xaa\xdb\x64\x6b\x77",
.klen = 32,
.iv = "\xae\xcc\xde\xd5\x07\x25\x1f\x91"
"\x39\x14\xa6\x0c\x55\xc4\x7b\x30",
.assoc = "",
.alen = 0,
.input = "\xcc\x5f\xfb\xa4\x8f\x63\x74\x9f"
"\x7a\x81\xff\x55\x52\x56\xdc\x33"
"\x01\x52\xcd\xdb\x53\x78\xd9\xe1"
"\xb7\x1d\x06\x8d\xff\xab\x22\x98",
.ilen = 32,
.result = "\xbb\x01\x7c\xd1\x2c\x33\x7b\x37"
"\x0a\xee\xc4\x30\x19\xd7\x3a\x6f"
"\xf8\x2b\x67\xf5\x3b\x84\x87\x2a"
"\xfb\x07\x7a\x82\xb5\xe4\x85\x26"
"\x1e\xa8\xe5\x04\x54\xce\xe5\x5f"
"\xb5\x3f\xc1\xd5\x7f\xbd\xd2\xa6",
.rlen = 48,
}, {
.key = "\xdb\x28\x89\x0c\xd3\x01\x1e\x26"
"\x9a\xb7\x2b\x79\x51\x1f\x0d\xb4"
"\xcc\x03\x50\xfc\x95\x20\xb9\xe7"
"\xe8\x29\x3e\x83\xb5\xc3\x0a\x4e",
.klen = 32,
.iv = "\xea\xf1\x18\x74\x17\xa0\xc8\xad"
"\xba\xee\x58\x9d\x4f\xe8\x3d\x36",
.assoc = "",
.alen = 0,
.input = "\x08\x84\x34\x44\x9f\xde\x1c\xbb"
"\xfb\x5b\xb1\xe6\x4c\x7a\x9f\x39"
"\x2c\x14\xd9\x5f\x59\x18\x5b\xfb"
"\x79\x4b\xe5\x65\xd9\x0a\xc1\x6f"
"\x2e",
.ilen = 33,
.result = "\xc2\xf4\x40\x55\xf9\x59\xff\x73"
"\x08\xf5\x98\x92\x0c\x7b\x35\x9a"
"\xa8\xf4\x42\x7e\x6f\x93\xca\x22"
"\x23\x06\x1e\xf8\x89\x22\xf4\x46"
"\x7c\x7c\x67\x75\xab\xe5\x75\xaa"
"\x15\xd7\x83\x19\xfd\x31\x59\x5b"
"\x32",
.rlen = 49,
}, {
.key = "\x17\x4d\xc3\xab\xe3\x7d\xc7\x42"
"\x1b\x91\xdd\x0a\x4b\x43\xcf\xba"
"\xf6\xc5\x5c\x80\x9a\xc0\x3b\x01"
"\xa9\x56\x1d\x5b\x8f\x22\xa9\x25",
.klen = 32,
.iv = "\x27\x16\x51\x13\x27\x1c\x71\xc9"
"\x3b\xc8\x0a\x2f\x49\x0c\x00\x3c",
.assoc = "",
.alen = 0,
.input = "\x45\xa8\x6e\xe3\xaf\x5a\xc5\xd7"
"\x7c\x35\x63\x77\x46\x9f\x61\x3f"
"\x56\xd7\xe4\xe3\x5e\xb8\xdc\x14"
"\x3a\x79\xc4\x3e\xb3\x69\x61\x46"
"\x3c\xb6\x83\x4e\xb4\x26\xc7\x73"
"\x22\xda\x52\x8b\x7d\x11\x98\xea"
"\x62\xe1\x14\x1e\xdc\xfe\x0f\xad"
"\x20\x76\x5a\xdc\x4e\x71\x13",
.ilen = 63,
.result = "\xc9\x82\x3b\x4b\x87\x84\xa5\xdb"
"\xa0\x8c\xd3\x3e\x7f\x8d\xe8\x28"
"\x2a\xdc\xfa\x01\x84\x87\x9a\x70"
"\x81\x75\x37\x0a\xd2\x75\xa9\xb6"
"\x21\x72\xee\x7e\x65\x95\xe5\xcc"
"\x01\xb7\x39\xa6\x51\x15\xca\xff"
"\x61\xdc\x97\x38\xcc\xf4\xca\xc7"
"\x83\x9b\x05\x11\x72\x60\xf0\xb4"
"\x7e\x06\xab\x0a\xc0\xbb\x59\x23"
"\xaa\x2d\xfc\x4e\x35\x05\x59",
.rlen = 79,
}, {
.key = "\x54\x71\xfd\x4b\xf3\xf9\x6f\x5e"
"\x9c\x6c\x8f\x9c\x45\x68\x92\xc1"
"\x21\x87\x67\x04\x9f\x60\xbd\x1b"
"\x6a\x84\xfc\x34\x6a\x81\x48\xfb",
.klen = 32,
.iv = "\x63\x3b\x8b\xb3\x37\x98\x1a\xe5"
"\xbc\xa2\xbc\xc0\x43\x31\xc2\x42",
.assoc = "",
.alen = 0,
.input = "\x81\xcd\xa8\x82\xbf\xd6\x6e\xf3"
"\xfd\x0f\x15\x09\x40\xc3\x24\x45"
"\x81\x99\xf0\x67\x63\x58\x5e\x2e"
"\xfb\xa6\xa3\x16\x8d\xc8\x00\x1c"
"\x4b\x62\x87\x7c\x15\x38\xda\x70"
"\x3d\xea\xe7\xf2\x40\xba\xae\x79"
"\x8f\x48\xfc\xbf\x45\x53\x2e\x78"
"\xef\x79\xf0\x1b\x49\xf7\xfd\x9c",
.ilen = 64,
.result = "\x11\x7c\x7d\xef\xce\x29\x95\xec"
"\x7e\x9f\x42\xa6\x26\x07\xa1\x75"
"\x2f\x4e\x09\x9a\xf6\x6b\xc2\xfa"
"\x0d\xd0\x17\xdc\x25\x1e\x9b\xdc"
"\x5f\x8c\x1c\x60\x15\x4f\x9b\x20"
"\x7b\xff\xcd\x82\x60\x84\xf4\xa5"
"\x20\x9a\x05\x19\x5b\x02\x0a\x72"
"\x43\x11\x26\x58\xcf\xc5\x41\xcf"
"\x13\xcc\xde\x32\x92\xfa\x86\xf2"
"\xaf\x16\xe8\x8f\xca\xb6\xfd\x54",
.rlen = 80,
}, {
.key = "\x90\x96\x36\xea\x03\x74\x18\x7a"
"\x1d\x46\x42\x2d\x3f\x8c\x54\xc7"
"\x4b\x4a\x73\x89\xa4\x00\x3f\x34"
"\x2c\xb1\xdb\x0c\x44\xe0\xe8\xd2",
.klen = 32,
.iv = "\xa0\x5f\xc5\x52\x47\x13\xc2\x01"
"\x3d\x7c\x6e\x52\x3d\x55\x85\x48",
.assoc = "\xaf",
.alen = 1,
.input = "",
.ilen = 0,
.result = "\x9b\xc5\x3b\x20\x0a\x88\x56\xbe"
"\x69\xdf\xc4\xc4\x02\x46\x3a\xf0",
.rlen = 16,
}, {
.key = "\xcd\xbb\x70\x89\x13\xf0\xc1\x95"
"\x9e\x20\xf4\xbf\x39\xb1\x17\xcd"
"\x76\x0c\x7f\x0d\xa9\xa0\xc1\x4e"
"\xed\xdf\xb9\xe4\x1e\x3f\x87\xa8",
.klen = 32,
.iv = "\xdc\x84\xfe\xf1\x58\x8f\x6b\x1c"
"\xbe\x57\x20\xe3\x37\x7a\x48\x4f",
.assoc = "\xeb\x4d\x8d\x59\x9c\x2e\x15\xa3"
"\xde\x8d\x4d\x07\x36\x43\x78\xd0"
"\x0b\x6d\x84\x4f\x2c\xf0\x82\x5b"
"\x4e\xf6\x29\xd1\x8b\x6f\x56",
.alen = 31,
.input = "",
.ilen = 0,
.result = "\xe0\x6d\xa1\x07\x98\x2f\x40\x2d"
"\x2e\x9a\xd6\x61\x43\xc0\x74\x69",
.rlen = 16,
}, {
.key = "\x0a\xe0\xaa\x29\x24\x6c\x6a\xb1"
"\x1f\xfa\xa6\x50\x33\xd5\xda\xd3"
"\xa0\xce\x8a\x91\xae\x40\x43\x68"
"\xae\x0d\x98\xbd\xf8\x9e\x26\x7f",
.klen = 32,
.iv = "\x19\xa9\x38\x91\x68\x0b\x14\x38"
"\x3f\x31\xd2\x74\x31\x9e\x0a\x55",
.assoc = "\x28\x72\xc7\xf8\xac\xaa\xbe\xbf"
"\x5f\x67\xff\x99\x30\x67\x3b\xd6"
"\x35\x2f\x90\xd3\x31\x90\x04\x74"
"\x0f\x23\x08\xa9\x65\xce\xf6\xea",
.alen = 32,
.input = "",
.ilen = 0,
.result = "\xb9\x57\x13\x3e\x82\x31\x61\x65"
"\x0d\x7f\x6c\x96\x93\x5c\x50\xe2",
.rlen = 16,
}, {
.key = "\x46\x04\xe3\xc8\x34\xe7\x12\xcd"
"\xa0\xd4\x58\xe2\x2d\xf9\x9c\xda"
"\xca\x91\x96\x15\xb4\xe0\xc5\x81"
"\x70\x3a\x77\x95\xd2\xfd\xc5\x55",
.klen = 32,
.iv = "\x55\xcd\x72\x30\x78\x86\xbd\x54"
"\xc0\x0b\x84\x06\x2b\xc2\xcd\x5b",
.assoc = "\x64\x97\x00\x98\xbc\x25\x67\xdb"
"\xe0\x41\xb1\x2a\x2a\x8c\xfe\xdd"
"\x5f\xf2\x9c\x58\x36\x30\x86\x8e"
"\xd1\x51\xe6\x81\x3f\x2d\x95\xc1"
"\x01",
.alen = 33,
.input = "",
.ilen = 0,
.result = "\x81\x96\x34\xde\xbb\x36\xdd\x3e"
"\x4e\x5e\xcb\x44\x21\xb8\x3f\xf1",
.rlen = 16,
}, {
.key = "\x83\x29\x1d\x67\x44\x63\xbb\xe9"
"\x20\xaf\x0a\x73\x27\x1e\x5f\xe0"
"\xf5\x53\xa1\x9a\xb9\x80\x47\x9b"
"\x31\x68\x56\x6e\xac\x5c\x65\x2c",
.klen = 32,
.iv = "\x92\xf2\xac\xcf\x88\x02\x65\x70"
"\x41\xe5\x36\x97\x25\xe7\x90\x61",
.assoc = "\xa1\xbb\x3a\x37\xcc\xa1\x10\xf7"
"\x61\x1c\x63\xbc\x24\xb0\xc0\xe3"
"\x8a\xb4\xa7\xdc\x3b\xd0\x08\xa8"
"\x92\x7f\xc5\x5a\x19\x8c\x34\x97"
"\x0f\x95\x9b\x18\xe4\x8d\xb4\x24"
"\xb9\x33\x28\x18\xe1\x9d\x14\xe0"
"\x64\xb2\x89\x7d\x78\xa8\x05\x7e"
"\x07\x8c\xfc\x88\x2d\xb8\x53",
.alen = 63,
.input = "",
.ilen = 0,
.result = "\x2e\x99\xb6\x79\x57\x56\x80\x36"
"\x8e\xc4\x1c\x12\x7d\x71\x36\x0c",
.rlen = 16,
}, {
.key = "\xbf\x4e\x57\x07\x54\xdf\x64\x05"
"\xa1\x89\xbc\x04\x21\x42\x22\xe6"
"\x1f\x15\xad\x1e\xbe\x20\xc9\xb4"
"\xf3\x95\x35\x46\x86\xbb\x04\x03",
.klen = 32,
.iv = "\xce\x17\xe5\x6f\x98\x7e\x0e\x8c"
"\xc2\xbf\xe8\x29\x1f\x0b\x52\x68",
.assoc = "\xdd\xe0\x74\xd6\xdc\x1d\xb8\x13"
"\xe2\xf6\x15\x4d\x1e\xd4\x83\xe9"
"\xb4\x76\xb3\x60\x40\x70\x8a\xc1"
"\x53\xac\xa4\x32\xf3\xeb\xd3\x6e"
"\x1e\x42\xa0\x46\x45\x9f\xc7\x22"
"\xd3\x43\xbc\x7e\xa5\x47\x2a\x6f"
"\x91\x19\x70\x1e\xe1\xfe\x25\x49"
"\xd6\x8f\x93\xc7\x28\x3f\x3d\x03",
.alen = 64,
.input = "",
.ilen = 0,
.result = "\x7b\x25\x3d\x47\xd4\xa7\x08\xce"
"\x3b\x89\x40\x36\xba\x6d\x0e\xa2",
.rlen = 16,
}, {
.key = "\xfc\x72\x90\xa6\x64\x5a\x0d\x21"
"\x22\x63\x6e\x96\x1b\x67\xe4\xec"
"\x49\xd7\xb9\xa2\xc3\xc0\x4b\xce"
"\xb4\xc3\x14\x1e\x61\x1a\xa3\xd9",
.klen = 32,
.iv = "\x0b\x3c\x1f\x0e\xa8\xf9\xb7\xa7"
"\x42\x9a\x9a\xba\x19\x30\x15\x6e",
.assoc = "\x1a",
.alen = 1,
.input = "\x29",
.ilen = 1,
.result = "\xe6\x09\x6f\x95\x9a\x18\xc8\xf6"
"\x17\x75\x81\x16\xdf\x26\xff\x67"
"\x92",
crypto: testmgr - add test vectors for partial ctr(cast5) and ctr(cast6) Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:54 -06:00
.rlen = 17,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
}, {
.key = "\x38\x97\xca\x45\x74\xd6\xb6\x3c"
"\xa3\x3d\x20\x27\x15\x8b\xa7\xf2"
"\x74\x9a\xc4\x27\xc8\x60\xcd\xe8"
"\x75\xf0\xf2\xf7\x3b\x79\x42\xb0",
.klen = 32,
.iv = "\x47\x60\x59\xad\xb8\x75\x60\xc3"
"\xc3\x74\x4c\x4c\x13\x54\xd8\x74",
.assoc = "\x56\x29\xe7\x15\xfc\x14\x0a\x4a"
"\xe4\xaa\x79\x70\x12\x1d\x08\xf6"
"\x09\xfb\xca\x69\x4b\xb0\x8e\xf5"
"\xd6\x07\x62\xe3\xa8\xa9\x12",
.alen = 31,
.input = "\x66\xf3\x75\x7d\x40\xb3\xb4\xd1"
"\x04\xe1\xa6\x94\x10\xe6\x39\x77"
"\xd3\xac\x4d\x8a\x8c\x58\x6e\xfb"
"\x06\x13\x9a\xd9\x5e\xc0\xfa",
.ilen = 31,
.result = "\x82\xc0\x56\xf0\xd7\xc4\xc9\xfd"
"\x3c\xd1\x2a\xd4\x15\x86\x9d\xda"
"\xea\x6c\x6f\xa1\x33\xb0\x7a\x01"
"\x57\xe7\xf3\x7b\x73\xe7\x54\x10"
"\xc6\x91\xe2\xc6\xa0\x69\xe7\xe6"
"\x76\xc3\xf5\x3a\x76\xfd\x4a",
.rlen = 47,
}, {
.key = "\x75\xbc\x04\xe5\x84\x52\x5e\x58"
"\x24\x17\xd2\xb9\x0e\xaf\x6a\xf9"
"\x9e\x5c\xd0\xab\xcd\x00\x4f\x01"
"\x37\x1e\xd1\xcf\x15\xd8\xe2\x86",
.klen = 32,
.iv = "\x84\x85\x92\x4d\xc8\xf1\x08\xdf"
"\x44\x4e\xff\xdd\x0d\x78\x9a\x7a",
.assoc = "\x93\x4e\x21\xb4\x0c\x90\xb3\x66"
"\x65\x84\x2b\x01\x0b\x42\xcb\xfc"
"\x33\xbd\xd6\xed\x50\x50\x10\x0e"
"\x97\x35\x41\xbb\x82\x08\xb1\xf2",
.alen = 32,
.input = "\xa2\x17\xaf\x1c\x50\x2e\x5d\xed"
"\x85\xbb\x58\x26\x0a\x0b\xfc\x7d"
"\xfe\x6e\x59\x0e\x91\xf8\xf0\x15"
"\xc8\x40\x78\xb1\x38\x1f\x99\xa7",
.ilen = 32,
.result = "\x01\x47\x8e\x6c\xf6\x64\x89\x3a"
"\x71\xce\xe4\xaa\x45\x70\xe6\x84"
"\x62\x48\x08\x64\x86\x6a\xdf\xec"
"\xb4\xa0\xfb\x34\x03\x0c\x19\xf4"
"\x2b\x7b\x36\x73\xec\x54\xa9\x1e"
"\x30\x85\xdb\xe4\xac\xe9\x2c\xca",
.rlen = 48,
}, {
.key = "\xb1\xe1\x3e\x84\x94\xcd\x07\x74"
"\xa5\xf2\x84\x4a\x08\xd4\x2c\xff"
"\xc8\x1e\xdb\x2f\xd2\xa0\xd1\x1b"
"\xf8\x4c\xb0\xa8\xef\x37\x81\x5d",
.klen = 32,
.iv = "\xc0\xaa\xcc\xec\xd8\x6c\xb1\xfb"
"\xc5\x28\xb1\x6e\x07\x9d\x5d\x81",
.assoc = "\xd0\x73\x5a\x54\x1d\x0b\x5b\x82"
"\xe5\x5f\xdd\x93\x05\x66\x8e\x02"
"\x5e\x80\xe1\x71\x55\xf0\x92\x28"
"\x59\x62\x20\x94\x5c\x67\x50\xc8"
"\x58",
.alen = 33,
.input = "\xdf\x3c\xe9\xbc\x61\xaa\x06\x09"
"\x06\x95\x0a\xb7\x04\x2f\xbe\x84"
"\x28\x30\x64\x92\x96\x98\x72\x2e"
"\x89\x6e\x57\x8a\x13\x7e\x38\x7e"
"\xdb",
.ilen = 33,
.result = "\x85\xe0\xf8\x0f\x8e\x49\xe3\x60"
"\xcb\x4a\x54\x94\xcf\xf5\x7e\x34"
"\xe9\xf8\x80\x65\x53\xd0\x72\x70"
"\x4f\x7d\x9d\xd1\x15\x6f\xb9\x2c"
"\xfa\xe8\xdd\xac\x2e\xe1\x3f\x67"
"\x63\x0f\x1a\x59\xb7\x89\xdb\xf4"
"\xc3",
.rlen = 49,
}, {
.key = "\xee\x05\x77\x23\xa5\x49\xb0\x90"
"\x26\xcc\x36\xdc\x02\xf8\xef\x05"
"\xf3\xe1\xe7\xb3\xd8\x40\x53\x35"
"\xb9\x79\x8f\x80\xc9\x96\x20\x33",
.klen = 32,
.iv = "\xfd\xce\x06\x8b\xe9\xe8\x5a\x17"
"\x46\x02\x63\x00\x01\xc1\x20\x87",
.assoc = "\x0c\x98\x94\xf3\x2d\x87\x04\x9e"
"\x66\x39\x8f\x24\xff\x8a\x50\x08"
"\x88\x42\xed\xf6\x5a\x90\x14\x42"
"\x1a\x90\xfe\x6c\x36\xc6\xf0\x9f"
"\x66\xa0\xb5\x2d\x2c\xf8\x25\x15"
"\x55\x90\xa2\x7e\x77\x94\x96\x3a"
"\x71\x1c\xf7\x44\xee\xa8\xc3\x42"
"\xe2\xa3\x84\x04\x0b\xe1\xce",
.alen = 63,
.input = "\x1b\x61\x23\x5b\x71\x26\xae\x25"
"\x87\x6f\xbc\x49\xfe\x53\x81\x8a"
"\x53\xf2\x70\x17\x9b\x38\xf4\x48"
"\x4b\x9b\x36\x62\xed\xdd\xd8\x54"
"\xea\xcb\xb6\x79\x45\xfc\xaa\x54"
"\x5c\x94\x47\x58\xa7\xff\x9c\x9e"
"\x7c\xb6\xf1\xac\xc8\xfd\x8b\x35"
"\xd5\xa4\x6a\xd4\x09\xc2\x08",
.ilen = 63,
.result = "\x00\xe5\x5b\x87\x5c\x20\x22\x8a"
"\xda\x1f\xd3\xff\xbb\xb2\xb0\xf8"
"\xef\xe9\xeb\x9e\x7c\x80\xf4\x2b"
"\x59\xc0\x79\xbc\x17\xa0\x15\x01"
"\xf5\x72\xfb\x5a\xe7\xaf\x07\xe3"
"\x1b\x49\x21\x34\x23\x63\x55\x5e"
"\xee\x4f\x34\x17\xfa\xfe\xa5\x0c"
"\xed\x0b\x23\xea\x9b\xda\x57\x2f"
"\xf6\xa9\xae\x0d\x4e\x40\x96\x45"
"\x7f\xfa\xf0\xbf\xc4\x98\x78",
.rlen = 79,
}, {
.key = "\x2a\x2a\xb1\xc3\xb5\xc5\x59\xac"
"\xa7\xa6\xe8\x6d\xfc\x1d\xb2\x0b"
"\x1d\xa3\xf3\x38\xdd\xe0\xd5\x4e"
"\x7b\xa7\x6e\x58\xa3\xf5\xbf\x0a",
.klen = 32,
.iv = "\x39\xf3\x3f\x2b\xf9\x64\x03\x33"
"\xc7\xdd\x15\x91\xfb\xe6\xe2\x8d",
.assoc = "\x49\xbc\xce\x92\x3d\x02\xad\xba"
"\xe7\x13\x41\xb6\xf9\xaf\x13\x0f"
"\xb2\x04\xf8\x7a\x5f\x30\x96\x5b"
"\xdc\xbd\xdd\x44\x10\x25\x8f\x75"
"\x75\x4d\xb9\x5b\x8e\x0a\x38\x13"
"\x6f\x9f\x36\xe4\x3a\x3e\xac\xc9"
"\x9d\x83\xde\xe5\x57\xfd\xe3\x0e"
"\xb1\xa7\x1b\x44\x05\x67\xb7\x37",
.alen = 64,
.input = "\x58\x85\x5c\xfa\x81\xa1\x57\x40"
"\x08\x4a\x6e\xda\xf8\x78\x44\x90"
"\x7d\xb5\x7b\x9b\xa1\xd8\x76\x62"
"\x0c\xc9\x15\x3b\xc7\x3c\x77\x2b"
"\xf8\x78\xba\xa7\xa6\x0e\xbd\x52"
"\x76\xa3\xdc\xbe\x6b\xa8\xb1\x2d"
"\xa9\x1d\xd8\x4e\x31\x53\xab\x00"
"\xa5\xa7\x01\x13\x04\x49\xf2\x04",
.ilen = 64,
.result = "\x28\xdd\xb9\x4a\x12\xc7\x0a\xe1"
"\x58\x06\x1a\x9b\x8c\x67\xdf\xeb"
"\x35\x35\x60\x9d\x06\x40\x65\xc1"
"\x93\xe8\xb3\x82\x50\x29\xdd\xb5"
"\x2b\xcb\xde\x18\x78\x6b\x42\xbe"
"\x6d\x24\xd0\xb2\x7d\xd7\x08\x8f"
"\x4a\x18\x98\xad\x8c\xf2\x97\xb4"
"\xf4\x77\xe4\xbf\x41\x3b\xc4\x06"
"\xce\x9e\x34\x81\xf0\x89\x11\x13"
"\x02\x65\xa1\x7c\xdf\x07\x33\x06",
.rlen = 80,
}, {
.key = "\x67\x4f\xeb\x62\xc5\x40\x01\xc7"
"\x28\x80\x9a\xfe\xf6\x41\x74\x12"
"\x48\x65\xfe\xbc\xe2\x80\x57\x68"
"\x3c\xd4\x4d\x31\x7d\x54\x5f\xe1",
.klen = 32,
.iv = "\x76\x18\x79\xca\x09\xdf\xac\x4e"
"\x48\xb7\xc7\x23\xf5\x0a\xa5\x93",
.assoc = "\x85\xe1\x08\x32\x4d\x7e\x56\xd5"
"\x68\xed\xf3\x47\xf3\xd3\xd6\x15"
"\xdd\xc7\x04\xfe\x64\xd0\x18\x75"
"\x9d\xeb\xbc\x1d\xea\x84\x2e\x4c"
"\x83\xf9\xbe\x8a\xef\x1c\x4b\x10"
"\x89\xaf\xcb\x4b\xfe\xe7\xc1\x58"
"\xca\xea\xc6\x87\xc0\x53\x03\xd9"
"\x80\xaa\xb2\x83\xff\xee\xa1\x6a"
"\x04",
.alen = 65,
.input = "\x94\xaa\x96\x9a\x91\x1d\x00\x5c"
"\x88\x24\x20\x6b\xf2\x9c\x06\x96"
"\xa7\x77\x87\x1f\xa6\x78\xf8\x7b"
"\xcd\xf6\xf4\x13\xa1\x9b\x16\x02"
"\x07\x24\xbf\xd5\x08\x20\xd0\x4f"
"\x90\xb3\x70\x24\x2f\x51\xc7\xbb"
"\xd6\x84\xc0\xef\x9a\xa8\xca\xcc"
"\x74\xab\x97\x53\xfe\xd0\xdb\x37"
"\x37\x6a\x0e\x9f\x3f\xa3\x2a\xe3"
"\x1b\x34\x6d\x51\x72\x2b\x17\xe7"
"\x4d\xaa\x2c\x18\xda\xa3\x33\x89"
"\x2a\x9f\xf4\xd2\xed\x76\x3d\x3f"
"\x3c\x15\x9d\x8e\x4f\x3c\x27\xb0"
"\x42\x3f\x2f\x8a\xd4\xc2\x10\xb2"
"\x27\x7f\xe3\x34\x80\x02\x49\x4b"
"\x07\x68\x22\x2a\x88\x25\x53\xb2"
"\x2f",
.ilen = 129,
.result = "\x85\x39\x69\x35\xfb\xf9\xb0\xa6"
"\x85\x43\x88\xd0\xd7\x78\x60\x19"
"\x3e\x1f\xb1\xa4\xd6\xc5\x96\xec"
"\xf7\x84\x85\xc7\x27\x0f\x74\x57"
"\x28\x9e\xdd\x90\x3c\x43\x12\xc5"
"\x51\x3d\x39\x8f\xa5\xf4\xe0\x0b"
"\x57\x04\xf1\x6d\xfe\x9b\x84\x27"
"\xe8\xeb\x4d\xda\x02\x0a\xc5\x49"
"\x1a\x55\x5e\x50\x56\x4d\x94\xda"
"\x20\xf8\x12\x54\x50\xb3\x11\xda"
"\xed\x44\x27\x67\xd5\xd1\x8b\x4b"
"\x38\x67\x56\x65\x59\xda\xe6\x97"
"\x81\xae\x2f\x92\x3b\xae\x22\x1c"
"\x91\x59\x38\x18\x00\xe8\xba\x92"
"\x04\x19\x56\xdf\xb0\x82\xeb\x6f"
"\x2e\xdb\x54\x3c\x4b\xbb\x60\x90"
"\x4c\x50\x10\x62\xba\x7a\xb1\x68"
"\x37\xd7\x87\x4e\xe4\x66\x09\x1f"
"\xa5",
.rlen = 145,
}, {
.key = "\xa3\x73\x24\x01\xd5\xbc\xaa\xe3"
"\xa9\x5a\x4c\x90\xf0\x65\x37\x18"
"\x72\x28\x0a\x40\xe7\x20\xd9\x82"
"\xfe\x02\x2b\x09\x57\xb3\xfe\xb7",
.klen = 32,
.iv = "\xb3\x3d\xb3\x69\x19\x5b\x54\x6a"
"\xc9\x91\x79\xb4\xef\x2e\x68\x99",
.assoc = "\xc2\x06\x41\xd1\x5d\xfa\xff\xf1"
"\xe9\xc7\xa5\xd9\xed\xf8\x98\x1b"
"\x07\x89\x10\x82\x6a\x70\x9a\x8f"
"\x5e\x19\x9b\xf5\xc5\xe3\xcd\x22"
"\x92\xa5\xc2\xb8\x51\x2e\x5e\x0e"
"\xa4\xbe\x5f\xb1\xc1\x90\xd7\xe7"
"\xf7\x52\xae\x28\x29\xa8\x22\xa4"
"\x4f\xae\x48\xc2\xfa\x75\x8b\x9e"
"\xce\x83\x2a\x88\x07\x55\xbb\x89"
"\xf6\xdf\xac\xdf\x83\x08\xbf\x7d"
"\xac\x30\x8b\x8e\x02\xac\x00\xf1"
"\x30\x46\xe1\xbc\x75\xbf\x49\xbb"
"\x26\x4e\x29\xf0\x2f\x21\xc6\x13"
"\x92\xd9\x3d\x11\xe4\x10\x00\x8e"
"\xd4\xd4\x58\x65\xa6\x2b\xe3\x25"
"\xb1\x8f\x15\x93\xe7\x71\xb9\x2c"
"\x4b",
.alen = 129,
.input = "\xd1\xcf\xd0\x39\xa1\x99\xa9\x78"
"\x09\xfe\xd2\xfd\xec\xc1\xc9\x9d"
"\xd2\x39\x93\xa3\xab\x18\x7a\x95"
"\x8f\x24\xd3\xeb\x7b\xfa\xb5\xd8"
"\x15\xd1\xc3\x04\x69\x32\xe3\x4d"
"\xaa\xc2\x04\x8b\xf2\xfa\xdc\x4a"
"\x02\xeb\xa8\x90\x03\xfd\xea\x97"
"\x43\xaf\x2e\x92\xf8\x57\xc5\x6a"
"\x00",
.ilen = 65,
.result = "\x7d\xde\x53\x22\xe4\x23\x3b\x30"
"\x78\xde\x35\x90\x7a\xd9\x0b\x93"
"\xf6\x0e\x0b\xed\x40\xee\x10\x9c"
"\x96\x3a\xd3\x34\xb2\xd0\x67\xcf"
"\x63\x7f\x2d\x0c\xcf\x96\xec\x64"
"\x1a\x87\xcc\x7d\x2c\x5e\x81\x4b"
"\xd2\x8f\x4c\x7c\x00\xb1\xb4\xe0"
"\x87\x4d\xb1\xbc\xd8\x78\x2c\x17"
"\xf2\x3b\xd8\x28\x40\xe2\x76\xf6"
"\x20\x13\x83\x46\xaf\xff\xe3\x0f"
"\x72",
.rlen = 81,
}, {
.key = "\xe0\x98\x5e\xa1\xe5\x38\x53\xff"
"\x2a\x35\xfe\x21\xea\x8a\xfa\x1e"
"\x9c\xea\x15\xc5\xec\xc0\x5b\x9b"
"\xbf\x2f\x0a\xe1\x32\x12\x9d\x8e",
.klen = 32,
.iv = "\xef\x61\xed\x08\x29\xd7\xfd\x86"
"\x4a\x6b\x2b\x46\xe9\x53\x2a\xa0",
.assoc = "\xfe\x2a\x7b\x70\x6d\x75\xa7\x0d"
"\x6a\xa2\x57\x6a\xe7\x1c\x5b\x21"
"\x31\x4b\x1b\x07\x6f\x10\x1c\xa8"
"\x20\x46\x7a\xce\x9f\x42\x6d\xf9",
.alen = 32,
.input = "\x0d\xf4\x09\xd8\xb1\x14\x51\x94"
"\x8a\xd8\x84\x8e\xe6\xe5\x8c\xa3"
"\xfc\xfc\x9e\x28\xb0\xb8\xfc\xaf"
"\x50\x52\xb1\xc4\x55\x59\x55\xaf",
.ilen = 32,
.result = "\x5a\xcd\x8c\x57\xf2\x6a\xb6\xbe"
"\x53\xc7\xaa\x9a\x60\x74\x9c\xc4"
"\xa2\xc2\xd0\x6d\xe1\x03\x63\xdc"
"\xbb\x51\x7e\x9c\x89\x73\xde\x4e"
"\x24\xf8\x52\x7c\x15\x41\x0e\xba"
"\x69\x0e\x36\x5f\x2f\x22\x8c",
.rlen = 47,
}, {
.key = "\x1c\xbd\x98\x40\xf5\xb3\xfc\x1b"
"\xaa\x0f\xb0\xb3\xe4\xae\xbc\x24"
"\xc7\xac\x21\x49\xf1\x60\xdd\xb5"
"\x80\x5d\xe9\xba\x0c\x71\x3c\x64",
.klen = 32,
.iv = "\x2c\x86\x26\xa8\x39\x52\xa6\xa2"
"\xcb\x45\xdd\xd7\xe3\x77\xed\xa6",
.assoc = "\x3b\x4f\xb5\x10\x7d\xf1\x50\x29"
"\xeb\x7c\x0a\xfb\xe1\x40\x1e\x27"
"\x5c\x0d\x27\x8b\x74\xb0\x9e\xc2"
"\xe1\x74\x59\xa6\x79\xa1\x0c\xd0",
.alen = 32,
.input = "\x4a\x18\x43\x77\xc1\x90\xfa\xb0"
"\x0b\xb2\x36\x20\xe0\x09\x4e\xa9"
"\x26\xbe\xaa\xac\xb5\x58\x7e\xc8"
"\x11\x7f\x90\x9c\x2f\xb8\xf4\x85",
.ilen = 32,
.result = "\x47\xd6\xce\x78\xd6\xbf\x4a\x51"
"\xb8\xda\x92\x3c\xfd\xda\xac\x8e"
"\x8d\x88\xd7\x4d\x90\xe5\xeb\xa1"
"\xab\xd6\x7c\x76\xad\xea\x7d\x76"
"\x53\xee\xb0\xcd\xd0\x02\xbb\x70"
"\x5b\x6f\x7b\xe2\x8c\xe8",
.rlen = 46,
}, {
.key = "\x59\xe1\xd2\xdf\x05\x2f\xa4\x37"
"\x2b\xe9\x63\x44\xde\xd3\x7f\x2b"
"\xf1\x6f\x2d\xcd\xf6\x00\x5f\xcf"
"\x42\x8a\xc8\x92\xe6\xd0\xdc\x3b",
.klen = 32,
.iv = "\x68\xab\x60\x47\x49\xce\x4f\xbe"
"\x4c\x20\x8f\x68\xdd\x9c\xb0\xac",
.assoc = "\x77\x74\xee\xaf\x8d\x6d\xf9\x45"
"\x6c\x56\xbc\x8d\xdb\x65\xe0\x2e"
"\x86\xd0\x32\x0f\x79\x50\x20\xdb"
"\xa2\xa1\x37\x7e\x53\x00\xab\xa6",
.alen = 32,
.input = "\x86\x3d\x7d\x17\xd1\x0c\xa3\xcc"
"\x8c\x8d\xe8\xb1\xda\x2e\x11\xaf"
"\x51\x80\xb5\x30\xba\xf8\x00\xe2"
"\xd3\xad\x6f\x75\x09\x18\x93\x5c",
.ilen = 32,
.result = "\x9f\xa9\x2b\xa4\x8f\x00\x05\x2b"
"\xe7\x68\x81\x51\xbb\xfb\xdf\x60"
"\xbb\xac\xe8\xc1\xdc\x68\xae\x68"
"\x3a\xcd\x7a\x06\x49\xfe\x80\x11"
"\xe6\x61\x99\xe2\xdd\xbe\x2c\xbf",
.rlen = 40,
}, {
.key = "\x96\x06\x0b\x7f\x15\xab\x4d\x53"
"\xac\xc3\x15\xd6\xd8\xf7\x42\x31"
"\x1b\x31\x38\x51\xfc\xa0\xe1\xe8"
"\x03\xb8\xa7\x6b\xc0\x2f\x7b\x11",
.klen = 32,
.iv = "\xa5\xcf\x9a\xe6\x59\x4a\xf7\xd9"
"\xcd\xfa\x41\xfa\xd7\xc0\x72\xb2",
.assoc = "\xb4\x99\x28\x4e\x9d\xe8\xa2\x60"
"\xed\x30\x6e\x1e\xd5\x89\xa3\x34"
"\xb1\x92\x3e\x93\x7e\xf0\xa2\xf5"
"\x64\xcf\x16\x57\x2d\x5f\x4a\x7d",
.alen = 32,
.input = "\xc3\x62\xb7\xb6\xe2\x87\x4c\xe7"
"\x0d\x67\x9a\x43\xd4\x52\xd4\xb5"
"\x7b\x43\xc1\xb5\xbf\x98\x82\xfc"
"\x94\xda\x4e\x4d\xe4\x77\x32\x32",
.ilen = 32,
.result = "\xe2\x34\xfa\x25\xfd\xfb\x89\x5e"
"\x5b\x4e\x0b\x15\x6e\x39\xfb\x0c"
"\x73\xc7\xd9\x6b\xbe\xce\x9b\x70"
"\xc7\x4f\x96\x16\x03\xfc\xea\xfb"
"\x56",
.rlen = 33,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct aead_testvec morus1280_dec_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.iv = "\x0f\xc9\x8e\x67\x44\x9e\xaa\x86"
"\x20\x36\x2c\x24\xfe\xc9\x30\x81",
.assoc = "",
.alen = 0,
.input = "\x91\x85\x0f\xf5\x52\x9e\xce\xce"
"\x65\x99\xc7\xbf\xd3\x76\xe8\x98",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3c\x24\x39\x9f\x10\x7b\xa8\x1b"
"\x80\xda\xb2\x91\xf9\x24\xc2\x06",
.klen = 16,
.iv = "\x4b\xed\xc8\x07\x54\x1a\x52\xa2"
"\xa1\x10\xde\xb5\xf8\xed\xf3\x87",
.assoc = "",
.alen = 0,
.input = "\x88\xc3\x4c\xf0\x2f\x43\x76\x13"
"\x96\xda\x76\x34\x33\x4e\xd5\x39"
"\x73",
.ilen = 17,
.result = "\x69",
.rlen = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x79\x49\x73\x3e\x20\xf7\x51\x37"
"\x01\xb4\x64\x22\xf3\x48\x85\x0c",
.klen = 16,
.iv = "\x88\x12\x01\xa6\x64\x96\xfb\xbe"
"\x22\xea\x90\x47\xf2\x11\xb5\x8e",
.assoc = "",
.alen = 0,
.input = "\x3e\x5c\x3b\x58\x3b\x7d\x2a\x22"
"\x75\x0b\x24\xa6\x0e\xc3\xde\x52"
"\x97\x0b\x64\xd4\xce\x90\x52\xf7"
"\xef\xdb\x6a\x38\xd2\xa8\xa1\x0d"
"\xe0\x61\x33\x24\xc6\x4d\x51\xbc"
"\xa4\x21\x74\xcf\x19\x16\x59",
.ilen = 47,
.result = "\xa6\xa4\x1e\x76\xec\xd4\x50\xcc"
"\x62\x58\xe9\x8f\xef\xa4\x17\x91"
"\xb4\x96\x9f\x6b\xce\x38\xa5\x46"
"\x13\x7d\x64\x93\xd7\x05\xf5",
.rlen = 31,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xb5\x6e\xad\xdd\x30\x72\xfa\x53"
"\x82\x8e\x16\xb4\xed\x6d\x47\x12",
.klen = 16,
.iv = "\xc4\x37\x3b\x45\x74\x11\xa4\xda"
"\xa2\xc5\x42\xd8\xec\x36\x78\x94",
.assoc = "",
.alen = 0,
.input = "\x30\x82\x9c\x2b\x67\xcb\xf9\x1f"
"\xde\x9f\x77\xb2\xda\x92\x61\x5c"
"\x09\x0b\x2d\x9a\x26\xaa\x1c\x06"
"\xab\x74\xb7\x2b\x95\x5f\x9f\xa1"
"\x9a\xff\x50\xa0\xa2\xff\xc5\xad"
"\x21\x8e\x84\x5c\x12\x61\xb2\xae",
.ilen = 48,
.result = "\xe2\xc9\x58\x15\xfc\x4f\xf8\xe8"
"\xe3\x32\x9b\x21\xe9\xc8\xd9\x97"
"\xde\x58\xab\xf0\xd3\xd8\x27\x60"
"\xd5\xaa\x43\x6b\xb1\x64\x95\xa4",
.rlen = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf2\x92\xe6\x7d\x40\xee\xa3\x6f"
"\x03\x68\xc8\x45\xe7\x91\x0a\x18",
.klen = 16,
.iv = "\x01\x5c\x75\xe5\x84\x8d\x4d\xf6"
"\x23\x9f\xf4\x6a\xe6\x5a\x3b\x9a",
.assoc = "",
.alen = 0,
.input = "\x67\x5d\x8e\x45\xc8\x39\xf5\x17"
"\xc1\x1d\x2a\xdd\x88\x67\xda\x1f"
"\x6d\xe8\x37\x28\x5a\xc1\x5e\x9f"
"\xa6\xec\xc6\x92\x05\x4b\xc0\xa3"
"\x63\xef\x88\xa4\x9b\x0a\x5c\xed"
"\x2b\x6a\xac\x63\x52\xaa\x10\x94"
"\xd0",
.ilen = 49,
.result = "\x1f\xee\x92\xb4\x0c\xcb\xa1\x04"
"\x64\x0c\x4d\xb2\xe3\xec\x9c\x9d"
"\x09\x1a\xb7\x74\xd8\x78\xa9\x79"
"\x96\xd8\x22\x43\x8c\xc3\x34\x7b"
"\xc4",
.rlen = 33,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x2e\xb7\x20\x1c\x50\x6a\x4b\x8b"
"\x84\x42\x7a\xd7\xe1\xb5\xcd\x1f",
.klen = 16,
.iv = "\x3d\x80\xae\x84\x94\x09\xf6\x12"
"\xa4\x79\xa6\xfb\xe0\x7f\xfd\xa0",
.assoc = "",
.alen = 0,
.input = "\x7d\x61\x1a\x35\x20\xcc\x07\x88"
"\x03\x98\x87\xcf\xc0\x6e\x4d\x19"
"\xe3\xd4\x0b\xfb\x29\x8f\x49\x1a"
"\x3a\x06\x77\xce\x71\x2c\xcd\xdd"
"\xed\xf6\xc9\xbe\xa6\x3b\xb8\xfc"
"\x6c\xbe\x77\xed\x74\x0e\x20\x85"
"\xd0\x65\xde\x24\x6f\xe3\x25\xc5"
"\xdf\x5b\x0f\xbd\x8a\x88\x78\xc9"
"\xe5\x81\x37\xde\x84\x7a\xf6\x84"
"\x99\x7a\x72\x9c\x54\x31\xa1",
.ilen = 79,
.result = "\x5c\x13\xcb\x54\x1c\x47\x4a\x1f"
"\xe5\xe6\xff\x44\xdd\x11\x5f\xa3"
"\x33\xdd\xc2\xf8\xdd\x18\x2b\x93"
"\x57\x05\x01\x1c\x66\x22\xd3\x51"
"\xd3\xdf\x18\xc9\x30\x66\xed\xb1"
"\x96\x58\xd5\x8c\x64\x8c\x7c\xf5"
"\x01\xd0\x74\x5f\x9b\xaa\xf6\xd1"
"\xe6\x16\xa2\xac\xde\x47\x40",
.rlen = 63,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x6b\xdc\x5a\xbb\x60\xe5\xf4\xa6"
"\x05\x1d\x2c\x68\xdb\xda\x8f\x25",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x7a\xa5\xe8\x23\xa4\x84\x9e\x2d"
"\x25\x53\x58\x8c\xda\xa3\xc0\xa6",
.assoc = "",
.alen = 0,
.input = "\x05\xc5\xb1\xf9\x1b\xb9\xab\x2c"
"\xa5\x07\x12\xa7\x12\x39\x60\x66"
"\x30\x81\x4a\x03\x78\x28\x45\x52"
"\xd2\x2b\x24\xfd\x8b\xa5\xb7\x66"
"\x6f\x45\xd7\x3b\x67\x6f\x51\xb9"
"\xc0\x3d\x6c\xca\x1e\xae\xff\xb6"
"\x79\xa9\xe4\x82\x5d\x4c\x2d\xdf"
"\xeb\x71\x40\xc9\x2c\x40\x45\x6d"
"\x73\x77\x01\xf3\x4f\xf3\x9d\x2a"
"\x5d\x57\xa8\xa1\x18\xa2\xad\xcb",
.ilen = 80,
.result = "\x98\x37\x05\xf3\x2c\xc2\xf3\x3b"
"\x66\xc0\xb1\xd5\xd7\x35\x21\xaa"
"\x5d\x9f\xce\x7c\xe2\xb8\xad\xad"
"\x19\x33\xe0\xf4\x40\x81\x72\x28"
"\xe1\x8b\x1c\xf8\x91\x78\xff\xaf"
"\xb0\x68\x69\xf2\x27\x35\x91\x84"
"\x2e\x37\x5b\x00\x04\xff\x16\x9c"
"\xb5\x19\x39\xeb\xd9\xcd\x29\x9a",
.rlen = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xa7\x00\x93\x5b\x70\x61\x9d\xc2"
"\x86\xf7\xde\xfa\xd5\xfe\x52\x2b",
.klen = 16,
.iv = "\xb6\xca\x22\xc3\xb4\x00\x47\x49"
"\xa6\x2d\x0a\x1e\xd4\xc7\x83\xad",
.assoc = "\xc5",
.alen = 1,
.input = "\x4d\xbf\x11\xac\x7f\x97\x0b\x2e"
"\x89\x3b\x9d\x0f\x83\x1c\x08\xc3",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe4\x25\xcd\xfa\x80\xdd\x46\xde"
"\x07\xd1\x90\x8b\xcf\x23\x15\x31",
.klen = 16,
.iv = "\xf3\xee\x5c\x62\xc4\x7c\xf0\x65"
"\x27\x08\xbd\xaf\xce\xec\x45\xb3",
.assoc = "\x02\xb8\xea\xca\x09\x1b\x9a\xec"
"\x47\x3e\xe9\xd4\xcc\xb5\x76\x34"
"\xe8\x73\x62\x64\xab\x50\xd0\xda"
"\x6b\x83\x66\xaf\x3e\x27\xc9",
.alen = 31,
.input = "\x5b\xc0\x8d\x54\xe4\xec\xbe\x38"
"\x03\x12\xf9\xcc\x9e\x46\x42\x92",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x20\x4a\x07\x99\x91\x58\xee\xfa"
"\x88\xab\x42\x1c\xc9\x47\xd7\x38",
.klen = 16,
.iv = "\x2f\x13\x95\x01\xd5\xf7\x99\x81"
"\xa8\xe2\x6f\x41\xc8\x10\x08\xb9",
.assoc = "\x3f\xdc\x24\x69\x19\x96\x43\x08"
"\xc8\x18\x9b\x65\xc6\xd9\x39\x3b"
"\x12\x35\x6e\xe8\xb0\xf0\x52\xf3"
"\x2d\xb0\x45\x87\x18\x86\x68\xf6",
.alen = 32,
.input = "\x48\xc5\xc3\x4c\x40\x2e\x2f\xc2"
"\x6d\x65\xe0\x67\x9c\x1d\xa0\xf0",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x5d\x6f\x41\x39\xa1\xd4\x97\x16"
"\x09\x85\xf4\xae\xc3\x6b\x9a\x3e",
.klen = 16,
.iv = "\x6c\x38\xcf\xa1\xe5\x73\x41\x9d"
"\x29\xbc\x21\xd2\xc2\x35\xcb\xbf",
.assoc = "\x7b\x01\x5d\x08\x29\x12\xec\x24"
"\x49\xf3\x4d\xf7\xc0\xfe\xfb\x41"
"\x3c\xf8\x79\x6c\xb6\x90\xd4\x0d"
"\xee\xde\x23\x60\xf2\xe5\x08\xcc"
"\x97",
.alen = 33,
.input = "\x28\x64\x78\x51\x55\xd8\x56\x4a"
"\x58\x3e\xf7\xbe\xee\x21\xfe\x94",
.ilen = 16,
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x99\x93\x7a\xd8\xb1\x50\x40\x31"
"\x8a\x60\xa6\x3f\xbd\x90\x5d\x44",
.klen = 16,
.iv = "\xa8\x5c\x09\x40\xf5\xef\xea\xb8"
"\xaa\x96\xd3\x64\xbc\x59\x8d\xc6",
.assoc = "\xb8\x26\x97\xa8\x39\x8e\x94\x3f"
"\xca\xcd\xff\x88\xba\x22\xbe\x47"
"\x67\xba\x85\xf1\xbb\x30\x56\x26"
"\xaf\x0b\x02\x38\xcc\x44\xa7\xa3"
"\xa6\xbf\x31\x93\x60\xcd\xda\x63"
"\x2c\xb1\xaa\x19\xc8\x19\xf8\xeb"
"\x03\xa1\xe8\xbe\x37\x54\xec\xa2"
"\xcd\x2c\x45\x58\xbd\x8e\x80",
.alen = 63,
.input = "\xb3\xa6\x00\x4e\x09\x20\xac\x21"
"\x77\x72\x69\x76\x2d\x36\xe5\xc8",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\xd6\xb8\xb4\x77\xc1\xcb\xe9\x4d"
"\x0a\x3a\x58\xd1\xb7\xb4\x1f\x4a",
.klen = 16,
.iv = "\xe5\x81\x42\xdf\x05\x6a\x93\xd4"
"\x2b\x70\x85\xf5\xb6\x7d\x50\xcc",
.assoc = "\xf4\x4a\xd1\x47\x49\x09\x3d\x5b"
"\x4b\xa7\xb1\x19\xb4\x46\x81\x4d"
"\x91\x7c\x91\x75\xc0\xd0\xd8\x40"
"\x71\x39\xe1\x10\xa6\xa3\x46\x7a"
"\xb4\x6b\x35\xc2\xc1\xdf\xed\x60"
"\x46\xc1\x3e\x7f\x8c\xc2\x0e\x7a"
"\x30\x08\xd0\x5f\xa0\xaa\x0c\x6d"
"\x9c\x2f\xdb\x97\xb8\x15\x69\x01",
.alen = 64,
.input = "\x65\x33\x7b\xa1\x63\xf4\x20\xdd"
"\xe4\xb9\x4a\xaa\x9a\x21\xaa\x14",
.ilen = 16,
.result = "",
.rlen = 0,
}, {
.key = "\x12\xdd\xee\x17\xd1\x47\x92\x69"
"\x8b\x14\x0a\x62\xb1\xd9\xe2\x50",
.klen = 16,
.iv = "\x22\xa6\x7c\x7f\x15\xe6\x3c\xf0"
"\xac\x4b\x37\x86\xb0\xa2\x13\xd2",
.assoc = "\x31",
.alen = 1,
.input = "\x1d\x47\x17\x34\x86\xf5\x54\x1a"
"\x6d\x28\xb8\x5d\x6c\xcf\xa0\xb9"
"\xbf",
.ilen = 17,
.result = "\x40",
.rlen = 1,
}, {
.key = "\x4f\x01\x27\xb6\xe1\xc3\x3a\x85"
"\x0c\xee\xbc\xf4\xab\xfd\xa5\x57",
.klen = 16,
.iv = "\x5e\xcb\xb6\x1e\x25\x62\xe4\x0c"
"\x2d\x25\xe9\x18\xaa\xc6\xd5\xd8",
.assoc = "\x6d\x94\x44\x86\x69\x00\x8f\x93"
"\x4d\x5b\x15\x3c\xa8\x8f\x06\x5a"
"\xe6\x01\xa8\x7e\xca\x10\xdc\x73"
"\xf4\x94\x9f\xc1\x5a\x61\x85",
.alen = 31,
.input = "\x78\x90\x52\xae\x0f\xf7\x2e\xef"
"\x63\x09\x08\x58\xb5\x56\xbd\x72"
"\x6e\x42\xcf\x27\x04\x7c\xdb\x92"
"\x18\xe9\xa4\x33\x90\xba\x62\xb5"
"\x70\xd3\x88\x9b\x4f\x05\xa7\x51"
"\x85\x87\x17\x09\x42\xed\x4e",
.ilen = 47,
.result = "\x7c\x5d\xd3\xee\xad\x9f\x39\x1a"
"\x6d\x92\x42\x61\xa7\x58\x37\xdb"
"\xb0\xb2\x2b\x9f\x0b\xb8\xbd\x7a"
"\x24\xa0\xd6\xb7\x11\x79\x6c",
.rlen = 31,
}, {
.key = "\x8b\x26\x61\x55\xf1\x3e\xe3\xa1"
"\x8d\xc8\x6e\x85\xa5\x21\x67\x5d",
.klen = 16,
.iv = "\x9b\xef\xf0\xbd\x35\xdd\x8d\x28"
"\xad\xff\x9b\xa9\xa4\xeb\x98\xdf",
.assoc = "\xaa\xb8\x7e\x25\x79\x7c\x37\xaf"
"\xce\x36\xc7\xce\xa2\xb4\xc9\x60"
"\x10\xc3\xb3\x02\xcf\xb0\x5e\x8d"
"\xb5\xc2\x7e\x9a\x35\xc0\x24\xfd",
.alen = 32,
.input = "\x1d\x2c\x57\xe0\x50\x38\x3d\x41"
"\x2e\x71\xc8\x3b\x92\x43\x58\xaf"
"\x5a\xfb\xad\x8f\xd9\xd5\x8a\x5e"
"\xdb\xf3\xcd\x3a\x2b\xe1\x2c\x1a"
"\xb0\xed\xe3\x0c\x6e\xf9\xf2\xd6"
"\x90\xe6\xb1\x0e\xa5\x8a\xac\xb7",
.ilen = 48,
.result = "\xb9\x82\x0c\x8d\xbd\x1b\xe2\x36"
"\xee\x6c\xf4\xf2\xa1\x7d\xf9\xe2"
"\xdb\x74\x36\x23\x11\x58\x3f\x93"
"\xe5\xcd\xb5\x90\xeb\xd8\x0c\xb3",
.rlen = 32,
}, {
.key = "\xc8\x4b\x9b\xf5\x01\xba\x8c\xbd"
"\x0e\xa3\x21\x16\x9f\x46\x2a\x63",
.klen = 16,
.iv = "\xd7\x14\x29\x5d\x45\x59\x36\x44"
"\x2e\xd9\x4d\x3b\x9e\x0f\x5b\xe5",
.assoc = "\xe6\xdd\xb8\xc4\x89\xf8\xe0\xca"
"\x4f\x10\x7a\x5f\x9c\xd8\x8b\x66"
"\x3b\x86\xbf\x86\xd4\x50\xe0\xa7"
"\x76\xef\x5c\x72\x0f\x1f\xc3\xd4"
"\xee",
.alen = 33,
.input = "\x59\x10\x84\x1c\x83\x4c\x8b\xfc"
"\xfd\x2e\x4b\x46\x84\xff\x78\x4e"
"\x50\xda\x5c\xb9\x61\x1d\xf5\xb9"
"\xfe\xbb\x7f\xae\x8c\xc1\x24\xbd"
"\x8c\x6f\x1f\x9b\xce\xc6\xc1\x37"
"\x08\x06\x5a\xe5\x96\x10\x95\xc2"
"\x5e",
.ilen = 49,
.result = "\xf5\xa6\x46\x2c\xce\x97\x8a\x51"
"\x6f\x46\xa6\x83\x9b\xa1\xbc\xe8"
"\x05\x36\x42\xa7\x16\xf8\xc1\xad"
"\xa7\xfb\x94\x68\xc5\x37\xab\x8a"
"\x72",
.rlen = 33,
}, {
.key = "\x05\x70\xd5\x94\x12\x36\x35\xd8"
"\x8f\x7d\xd3\xa8\x99\x6a\xed\x69",
.klen = 16,
.iv = "\x14\x39\x63\xfc\x56\xd5\xdf\x5f"
"\xaf\xb3\xff\xcc\x98\x33\x1d\xeb",
.assoc = "\x23\x02\xf1\x64\x9a\x73\x89\xe6"
"\xd0\xea\x2c\xf1\x96\xfc\x4e\x6d"
"\x65\x48\xcb\x0a\xda\xf0\x62\xc0"
"\x38\x1d\x3b\x4a\xe9\x7e\x62\xaa"
"\xfd\xc9\x4a\xa9\xa9\x39\x4b\x54"
"\xc8\x0e\x24\x7f\x5e\x10\x7a\x45"
"\x10\x0b\x56\x85\xad\x54\xaa\x66"
"\xa8\x43\xcd\xd4\x9b\xb7\xfa",
.alen = 63,
.input = "\x9a\x12\xbc\xdf\x72\xa8\x56\x22"
"\x49\x2d\x07\x92\xfc\x3d\x6d\x5f"
"\xef\x36\x19\xae\x91\xfa\xd6\x63"
"\x46\xea\x8a\x39\x14\x21\xa6\x37"
"\x18\xfc\x97\x3e\x16\xa5\x4d\x39"
"\x45\x2e\x69\xcc\x9c\x5f\xdf\x6d"
"\x5e\xa2\xbf\xac\x83\x32\x72\x52"
"\x58\x58\x23\x40\xfd\xa5\xc2\xe6"
"\xe9\x5a\x50\x98\x00\x58\xc9\x86"
"\x4f\x20\x37\xdb\x7b\x22\xa3",
.ilen = 79,
.result = "\x32\xcb\x80\xcc\xde\x12\x33\x6d"
"\xf0\x20\x58\x15\x95\xc6\x7f\xee"
"\x2f\xf9\x4e\x2c\x1b\x98\x43\xc7"
"\x68\x28\x73\x40\x9f\x96\x4a\x60"
"\x80\xf4\x4b\xf4\xc1\x3d\xd0\x93"
"\xcf\x12\xc9\x59\x8f\x7a\x7f\xa8"
"\x1b\xa5\x50\xed\x87\xa9\x72\x59"
"\x9c\x44\xb2\xa4\x99\x98\x34",
.rlen = 63,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x41\x94\x0e\x33\x22\xb1\xdd\xf4"
"\x10\x57\x85\x39\x93\x8f\xaf\x70",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x50\x5d\x9d\x9b\x66\x50\x88\x7b"
"\x30\x8e\xb1\x5e\x92\x58\xe0\xf1",
.assoc = "\x5f\x27\x2b\x03\xaa\xef\x32\x02"
"\x50\xc4\xde\x82\x90\x21\x11\x73"
"\x8f\x0a\xd6\x8f\xdf\x90\xe4\xda"
"\xf9\x4a\x1a\x23\xc3\xdd\x02\x81"
"\x0b\x76\x4f\xd7\x0a\x4b\x5e\x51"
"\xe3\x1d\xb9\xe5\x21\xb9\x8f\xd4"
"\x3d\x72\x3e\x26\x16\xa9\xca\x32"
"\x77\x47\x63\x14\x95\x3d\xe4\x34",
.alen = 64,
.input = "\xe6\xeb\x92\x5a\x5b\xf0\x2d\xbb"
"\x23\xec\x35\xe3\xae\xc9\xfb\x0b"
"\x90\x14\x46\xeb\xa8\x8d\xb0\x9b"
"\x39\xda\x8b\x48\xec\xb2\x00\x4e"
"\x80\x6f\x46\x4f\x9b\x1e\xbb\x35"
"\xea\x5a\xbc\xa2\x36\xa5\x89\x45"
"\xc2\xd6\xd7\x15\x0b\xf6\x6c\x56"
"\xec\x99\x7d\x61\xb3\x15\x93\xed"
"\x83\x1e\xd9\x48\x84\x0b\x37\xfe"
"\x95\x74\x44\xd5\x54\xa6\x27\x06",
.ilen = 80,
.result = "\x6e\xf0\xba\x6b\xee\x8e\xdc\x89"
"\x71\xfb\x0a\xa6\x8f\xea\x41\xf4"
"\x5a\xbb\x59\xb0\x20\x38\xc5\xe0"
"\x29\x56\x52\x19\x79\xf5\xe9\x37"
"\x8f\xa1\x50\x23\x22\x4f\xe3\x91"
"\xe9\x21\x5e\xbf\x52\x23\x95\x37"
"\x48\x0c\x38\x8f\xf0\xff\x92\x24"
"\x6b\x47\x49\xe3\x94\x1f\x1e\x01",
.rlen = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x7e\xb9\x48\xd3\x32\x2d\x86\x10"
"\x91\x31\x37\xcb\x8d\xb3\x72\x76",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x8d\x82\xd6\x3b\x76\xcc\x30\x97"
"\xb1\x68\x63\xef\x8c\x7c\xa3\xf7",
.assoc = "\x9c\x4b\x65\xa2\xba\x6b\xdb\x1e"
"\xd1\x9e\x90\x13\x8a\x45\xd3\x79"
"\xba\xcd\xe2\x13\xe4\x30\x66\xf4"
"\xba\x78\xf9\xfb\x9d\x3c\xa1\x58"
"\x1a\x22\x53\x05\x6b\x5c\x71\x4f"
"\xfd\x2d\x4d\x4c\xe5\x62\xa5\x63"
"\x6a\xda\x26\xc8\x7f\xff\xea\xfd"
"\x46\x4a\xfa\x53\x8f\xc4\xcd\x68"
"\x58",
.alen = 65,
.input = "\x89\x24\x27\x86\xdc\xd7\x6b\xd9"
"\xd1\xcd\xdc\x16\xdd\x2c\xc1\xfb"
"\x52\xb5\xb3\xab\x50\x99\x3f\xa0"
"\x38\xa4\x74\xa5\x04\x15\x63\x05"
"\x8f\x54\x81\x06\x5a\x6b\xa4\x63"
"\x6d\xa7\x21\xcb\xff\x42\x30\x8e"
"\x3b\xd1\xca\x3f\x4b\x1a\xb8\xc3"
"\x42\x01\xe6\xbc\x75\x15\x87\xee"
"\xc9\x8e\x65\x01\xd9\xd8\xb5\x9f"
"\x48\x86\xa6\x5f\x2c\xc7\xb5\xb0"
"\xed\x5d\x14\x7c\x3f\x40\xb1\x0b"
"\x72\xef\x94\x8d\x7a\x85\x56\xe5"
"\x56\x08\x15\x56\xba\xaf\xbd\xf0"
"\x20\xef\xa0\xf6\xa9\xad\xa2\xc9"
"\x1c\x3b\x28\x51\x7e\x77\xb2\x18"
"\x4f\x61\x64\x37\x22\x36\x6d\x78"
"\xed\xed\x35\xe8\x83\xa5\xec\x25"
"\x6b\xff\x5f\x1a\x09\x96\x3d\xdc"
"\x20",
.ilen = 145,
.result = "\xab\x14\xf3\x0a\xfe\x0a\x85\xa5"
"\xf2\xd5\xbc\x38\x89\x0e\x04\xfb"
"\x84\x7d\x65\x34\x25\xd8\x47\xfa"
"\xeb\x83\x31\xf1\x54\x54\x89\x0d"
"\x9d\x4d\x54\x51\x84\x61\xf6\x8e"
"\x03\x31\xf2\x25\x16\xcc\xaa\xc6"
"\x75\x73\x20\x30\x59\x54\xb2\xf0"
"\x3a\x4b\xe0\x23\x8e\xa6\x08\x35"
"\x8a\xdf\x27\xa0\xe4\x60\x99\xae"
"\x8e\x43\xd9\x39\x7b\x10\x40\x67"
"\x5c\x7e\xc9\x70\x63\x34\xca\x59"
"\xfe\x86\xbc\xb7\x9c\x39\xf3\x6d"
"\x6a\x41\x64\x6f\x16\x7f\x65\x7e"
"\x89\x84\x68\xeb\xb0\x51\xbe\x55"
"\x33\x16\x59\x6c\x3b\xef\x88\xad"
"\x2f\xab\xbc\x25\x76\x87\x41\x2f"
"\x36",
.rlen = 129,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xba\xde\x82\x72\x42\xa9\x2f\x2c"
"\x12\x0b\xe9\x5c\x87\xd7\x35\x7c",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xc9\xa7\x10\xda\x86\x48\xd9\xb3"
"\x32\x42\x15\x80\x85\xa1\x65\xfe",
.assoc = "\xd8\x70\x9f\x42\xca\xe6\x83\x3a"
"\x52\x79\x42\xa5\x84\x6a\x96\x7f"
"\xe4\x8f\xed\x97\xe9\xd0\xe8\x0d"
"\x7c\xa6\xd8\xd4\x77\x9b\x40\x2e"
"\x28\xce\x57\x34\xcd\x6e\x84\x4c"
"\x17\x3c\xe1\xb2\xa8\x0b\xbb\xf1"
"\x96\x41\x0d\x69\xe8\x54\x0a\xc8"
"\x15\x4e\x91\x92\x89\x4b\xb7\x9b"
"\x21\xf7\x42\x89\xac\x12\x2a\x54"
"\x69\xee\x18\xc7\x8d\xed\xe8\xfd"
"\xbb\x04\x28\xe6\x8a\x3c\x98\xc1"
"\x04\x2d\xa9\xa1\x24\x83\xff\xe9"
"\x55\x7a\xf0\xd1\xf6\x63\x05\xe1"
"\xd9\x1e\x75\x72\xc1\x9f\xae\x32"
"\xe1\x6b\xcd\x9e\x61\x19\x23\x86"
"\xd9\xd2\xaf\x8e\xd5\xd3\xa8\xa9"
"\x51",
.alen = 129,
.input = "\x36\x78\xb9\x22\xde\x62\x35\x55"
"\x1a\x7a\xf5\x45\xbc\xd7\x15\x82"
"\x01\xe9\x5a\x07\xea\x46\xaf\x91"
"\xcb\x73\xa5\xee\xe1\xb4\xbf\xc2"
"\xdb\xd2\x9d\x59\xde\xfc\x83\x00"
"\xf5\x46\xac\x97\xd5\x57\xa9\xb9"
"\x1f\x8c\xe8\xca\x68\x8b\x91\x0c"
"\x01\xbe\x0a\xaf\x7c\xf6\x67\xa4"
"\xbf\xbc\x88\x3f\x5d\xd1\xf9\x19"
"\x0f\x9d\xb2\xaf\xb9\x6e\x17\xdf"
"\xa2",
.ilen = 81,
.result = "\xe8\x39\x2d\xaa\x0e\x85\x2d\xc1"
"\x72\xaf\x6e\xc9\x82\x33\xc7\x01"
"\xaf\x40\x70\xb8\x2a\x78\xc9\x14"
"\xac\xb1\x10\xca\x2e\xb3\x28\xe4"
"\xac\xfa\x58\x7f\xe5\x73\x09\x8c"
"\x1d\x40\x87\x8c\xd9\x75\xc0\x55"
"\xa2\xda\x07\xd1\xc2\xa9\xd1\xbb"
"\x09\x4f\x77\x62\x88\x2d\xf2\x68"
"\x54",
.rlen = 65,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf7\x02\xbb\x11\x52\x24\xd8\x48"
"\x93\xe6\x9b\xee\x81\xfc\xf7\x82",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x06\xcc\x4a\x79\x96\xc3\x82\xcf"
"\xb3\x1c\xc7\x12\x7f\xc5\x28\x04",
.assoc = "\x15\x95\xd8\xe1\xda\x62\x2c\x56"
"\xd3\x53\xf4\x36\x7e\x8e\x59\x85"
"\x0e\x51\xf9\x1c\xee\x70\x6a\x27"
"\x3d\xd3\xb7\xac\x51\xfa\xdf\x05",
.alen = 32,
.input = "\x08\x1b\x95\x0e\x41\x95\x02\x4b"
"\x9c\xbb\xa8\xd0\x7c\xd3\x44\x6e"
"\x89\x14\x33\x70\x0a\xbc\xea\x39"
"\x88\xaa\x2b\xd5\x73\x11\x55\xf5"
"\x33\x33\x9c\xd7\x42\x34\x49\x8e"
"\x2f\x03\x30\x05\x47\xaf\x34",
.ilen = 47,
.result = "\x24\x5e\x67\x49\x1e\x01\xd6\xdd"
"\xf3\x89\x20\x5b\x7c\x57\x89\x07"
"\xd9\x02\x7c\x3d\x2f\x18\x4b\x2d"
"\x6e\xde\xee\xa2\x08\x12\xc7\xba",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x33\x27\xf5\xb1\x62\xa0\x80\x63"
"\x14\xc0\x4d\x7f\x7b\x20\xba\x89",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x42\xf0\x84\x19\xa6\x3f\x2b\xea"
"\x34\xf6\x79\xa3\x79\xe9\xeb\x0a",
.assoc = "\x51\xb9\x12\x80\xea\xde\xd5\x71"
"\x54\x2d\xa6\xc8\x78\xb2\x1b\x8c"
"\x39\x14\x05\xa0\xf3\x10\xec\x41"
"\xff\x01\x95\x84\x2b\x59\x7f\xdb",
.alen = 32,
.input = "\x97\xca\xf4\xe0\x8d\x89\xbf\x68"
"\x0c\x60\xb9\x27\xdf\xaa\x41\xc6"
"\x25\xd8\xf7\x1f\x10\x15\x48\x61"
"\x4c\x95\x00\xdf\x51\x9b\x7f\xe6"
"\x24\x40\x9e\xbe\x3b\xeb\x1b\x98"
"\xb9\x9c\xe5\xef\xf2\x05",
.ilen = 46,
.result = "\x61\x83\xa0\xe8\x2e\x7d\x7f\xf8"
"\x74\x63\xd2\xec\x76\x7c\x4c\x0d"
"\x03\xc4\x88\xc1\x35\xb8\xcd\x47"
"\x2f\x0c\xcd\x7a\xe2\x71\x66\x91",
.rlen = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x70\x4c\x2f\x50\x72\x1c\x29\x7f"
"\x95\x9a\xff\x10\x75\x45\x7d\x8f",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x7f\x15\xbd\xb8\xb6\xba\xd3\x06"
"\xb5\xd1\x2b\x35\x73\x0e\xad\x10",
.assoc = "\x8e\xde\x4c\x20\xfa\x59\x7e\x8d"
"\xd5\x07\x58\x59\x72\xd7\xde\x92"
"\x63\xd6\x10\x24\xf8\xb0\x6e\x5a"
"\xc0\x2e\x74\x5d\x06\xb8\x1e\xb2",
.alen = 32,
.input = "\x63\x4c\x2a\x8e\xb4\x6b\x63\x0d"
"\xb5\xec\x9b\x4e\x12\x23\xa3\xcf"
"\x1a\x5a\x70\x15\x5a\x10\x40\x51"
"\xca\x47\x4c\x9d\xc9\x97\xf4\x77"
"\xdb\xc8\x10\x2d\xdc\x65\x20\x3f",
.ilen = 40,
.result = "\x9d\xa7\xda\x88\x3e\xf8\x28\x14"
"\xf5\x3e\x85\x7d\x70\xa0\x0f\x13"
"\x2e\x86\x93\x45\x3a\x58\x4f\x61"
"\xf0\x3a\xac\x53\xbc\xd0\x06\x68",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xac\x70\x69\xef\x82\x97\xd2\x9b"
"\x15\x74\xb1\xa2\x6f\x69\x3f\x95",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xbb\x3a\xf7\x57\xc6\x36\x7c\x22"
"\x36\xab\xde\xc6\x6d\x32\x70\x17",
.assoc = "\xcb\x03\x85\xbf\x0a\xd5\x26\xa9"
"\x56\xe1\x0a\xeb\x6c\xfb\xa1\x98"
"\x8d\x98\x1c\xa8\xfe\x50\xf0\x74"
"\x81\x5c\x53\x35\xe0\x17\xbd\x88",
.alen = 32,
.input = "\xf1\x62\x44\xc7\x5f\x19\xca\x43"
"\x47\x2c\xaf\x68\x82\xbd\x51\xef"
"\x3d\x65\xd8\x45\x2d\x06\x07\x78"
"\x08\x2e\xb3\x23\xcd\x81\x12\x55"
"\x1a",
.ilen = 33,
.result = "\xda\xcc\x14\x27\x4e\x74\xd1\x30"
"\x76\x18\x37\x0f\x6a\xc4\xd1\x1a"
"\x58\x49\x9f\xc9\x3f\xf8\xd1\x7a"
"\xb2\x67\x8b\x2b\x96\x2f\xa5\x3e",
.rlen = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xe9\x95\xa2\x8f\x93\x13\x7b\xb7"
"\x96\x4e\x63\x33\x69\x8d\x02\x9b"
"\x23\xf9\x22\xeb\x80\xa0\xb1\x81"
"\xe2\x73\xc3\x21\x4d\x47\x8d\xf4",
.klen = 32,
.iv = "\xf8\x5e\x31\xf7\xd7\xb2\x25\x3e"
"\xb7\x85\x90\x58\x67\x57\x33\x1d",
.assoc = "",
.alen = 0,
.input = "\xdf\x2f\x83\xc0\x45\x4a\x2c\xcf"
"\xb9\xd2\x41\xf6\x80\xa1\x52\x70",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x25\xba\xdc\x2e\xa3\x8f\x24\xd3"
"\x17\x29\x15\xc5\x63\xb2\xc5\xa1"
"\x4d\xbc\x2d\x6f\x85\x40\x33\x9a"
"\xa3\xa0\xa1\xfa\x27\xa6\x2c\xca",
.klen = 32,
.iv = "\x34\x83\x6a\x96\xe7\x2d\xce\x5a"
"\x38\x5f\x42\xe9\x61\x7b\xf5\x23",
.assoc = "",
.alen = 0,
.input = "\x01\xd8\x55\x3c\xc0\x5a\x4b\xc7"
"\x01\xf4\x08\xe3\x0d\xf7\xf0\x78"
"\x53",
.ilen = 17,
.result = "\x53",
.rlen = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x62\xdf\x16\xcd\xb3\x0a\xcc\xef"
"\x98\x03\xc7\x56\x5d\xd6\x87\xa8"
"\x77\x7e\x39\xf3\x8a\xe0\xb5\xb4"
"\x65\xce\x80\xd2\x01\x05\xcb\xa1",
.klen = 32,
.iv = "\x71\xa8\xa4\x35\xf7\xa9\x76\x75"
"\xb8\x39\xf4\x7a\x5b\x9f\xb8\x29",
.assoc = "",
.alen = 0,
.input = "\xc2\x4b\x41\x0f\x2d\xb9\x62\x07"
"\xff\x8e\x74\xf8\xa1\xa6\xd5\x37"
"\xa5\x64\x31\x5c\xca\x73\x9b\x43"
"\xe6\x70\x63\x46\x95\xcb\xf7\xb5"
"\x20\x8c\x75\x7a\x2a\x17\x2f\xa9"
"\xb8\x4d\x11\x42\xd1\xf8\xf1",
.ilen = 47,
.result = "\x8f\x3a\xc1\x05\x7f\xe7\xcb\x83"
"\xf9\xa6\x4d\xc3\x58\x31\x19\x2c"
"\xd7\x90\xc2\x56\x4e\xd8\x57\xc7"
"\xf6\xf0\x27\xb4\x25\x4c\x83",
.rlen = 31,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x9e\x03\x4f\x6d\xc3\x86\x75\x0a"
"\x19\xdd\x79\xe8\x57\xfb\x4a\xae"
"\xa2\x40\x45\x77\x90\x80\x37\xce"
"\x26\xfb\x5f\xaa\xdb\x64\x6b\x77",
.klen = 32,
.iv = "\xae\xcc\xde\xd5\x07\x25\x1f\x91"
"\x39\x14\xa6\x0c\x55\xc4\x7b\x30",
.assoc = "",
.alen = 0,
.input = "\xbb\x01\x7c\xd1\x2c\x33\x7b\x37"
"\x0a\xee\xc4\x30\x19\xd7\x3a\x6f"
"\xf8\x2b\x67\xf5\x3b\x84\x87\x2a"
"\xfb\x07\x7a\x82\xb5\xe4\x85\x26"
"\x1e\xa8\xe5\x04\x54\xce\xe5\x5f"
"\xb5\x3f\xc1\xd5\x7f\xbd\xd2\xa6",
.ilen = 48,
.result = "\xcc\x5f\xfb\xa4\x8f\x63\x74\x9f"
"\x7a\x81\xff\x55\x52\x56\xdc\x33"
"\x01\x52\xcd\xdb\x53\x78\xd9\xe1"
"\xb7\x1d\x06\x8d\xff\xab\x22\x98",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xdb\x28\x89\x0c\xd3\x01\x1e\x26"
"\x9a\xb7\x2b\x79\x51\x1f\x0d\xb4"
"\xcc\x03\x50\xfc\x95\x20\xb9\xe7"
"\xe8\x29\x3e\x83\xb5\xc3\x0a\x4e",
.klen = 32,
.iv = "\xea\xf1\x18\x74\x17\xa0\xc8\xad"
"\xba\xee\x58\x9d\x4f\xe8\x3d\x36",
.assoc = "",
.alen = 0,
.input = "\xc2\xf4\x40\x55\xf9\x59\xff\x73"
"\x08\xf5\x98\x92\x0c\x7b\x35\x9a"
"\xa8\xf4\x42\x7e\x6f\x93\xca\x22"
"\x23\x06\x1e\xf8\x89\x22\xf4\x46"
"\x7c\x7c\x67\x75\xab\xe5\x75\xaa"
"\x15\xd7\x83\x19\xfd\x31\x59\x5b"
"\x32",
.ilen = 49,
.result = "\x08\x84\x34\x44\x9f\xde\x1c\xbb"
"\xfb\x5b\xb1\xe6\x4c\x7a\x9f\x39"
"\x2c\x14\xd9\x5f\x59\x18\x5b\xfb"
"\x79\x4b\xe5\x65\xd9\x0a\xc1\x6f"
"\x2e",
.rlen = 33,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x17\x4d\xc3\xab\xe3\x7d\xc7\x42"
"\x1b\x91\xdd\x0a\x4b\x43\xcf\xba"
"\xf6\xc5\x5c\x80\x9a\xc0\x3b\x01"
"\xa9\x56\x1d\x5b\x8f\x22\xa9\x25",
.klen = 32,
.iv = "\x27\x16\x51\x13\x27\x1c\x71\xc9"
"\x3b\xc8\x0a\x2f\x49\x0c\x00\x3c",
.assoc = "",
.alen = 0,
.input = "\xc9\x82\x3b\x4b\x87\x84\xa5\xdb"
"\xa0\x8c\xd3\x3e\x7f\x8d\xe8\x28"
"\x2a\xdc\xfa\x01\x84\x87\x9a\x70"
"\x81\x75\x37\x0a\xd2\x75\xa9\xb6"
"\x21\x72\xee\x7e\x65\x95\xe5\xcc"
"\x01\xb7\x39\xa6\x51\x15\xca\xff"
"\x61\xdc\x97\x38\xcc\xf4\xca\xc7"
"\x83\x9b\x05\x11\x72\x60\xf0\xb4"
"\x7e\x06\xab\x0a\xc0\xbb\x59\x23"
"\xaa\x2d\xfc\x4e\x35\x05\x59",
.ilen = 79,
.result = "\x45\xa8\x6e\xe3\xaf\x5a\xc5\xd7"
"\x7c\x35\x63\x77\x46\x9f\x61\x3f"
"\x56\xd7\xe4\xe3\x5e\xb8\xdc\x14"
"\x3a\x79\xc4\x3e\xb3\x69\x61\x46"
"\x3c\xb6\x83\x4e\xb4\x26\xc7\x73"
"\x22\xda\x52\x8b\x7d\x11\x98\xea"
"\x62\xe1\x14\x1e\xdc\xfe\x0f\xad"
"\x20\x76\x5a\xdc\x4e\x71\x13",
.rlen = 63,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x54\x71\xfd\x4b\xf3\xf9\x6f\x5e"
"\x9c\x6c\x8f\x9c\x45\x68\x92\xc1"
"\x21\x87\x67\x04\x9f\x60\xbd\x1b"
"\x6a\x84\xfc\x34\x6a\x81\x48\xfb",
.klen = 32,
.iv = "\x63\x3b\x8b\xb3\x37\x98\x1a\xe5"
"\xbc\xa2\xbc\xc0\x43\x31\xc2\x42",
.assoc = "",
.alen = 0,
.input = "\x11\x7c\x7d\xef\xce\x29\x95\xec"
"\x7e\x9f\x42\xa6\x26\x07\xa1\x75"
"\x2f\x4e\x09\x9a\xf6\x6b\xc2\xfa"
"\x0d\xd0\x17\xdc\x25\x1e\x9b\xdc"
"\x5f\x8c\x1c\x60\x15\x4f\x9b\x20"
"\x7b\xff\xcd\x82\x60\x84\xf4\xa5"
"\x20\x9a\x05\x19\x5b\x02\x0a\x72"
"\x43\x11\x26\x58\xcf\xc5\x41\xcf"
"\x13\xcc\xde\x32\x92\xfa\x86\xf2"
"\xaf\x16\xe8\x8f\xca\xb6\xfd\x54",
.ilen = 80,
.result = "\x81\xcd\xa8\x82\xbf\xd6\x6e\xf3"
"\xfd\x0f\x15\x09\x40\xc3\x24\x45"
"\x81\x99\xf0\x67\x63\x58\x5e\x2e"
"\xfb\xa6\xa3\x16\x8d\xc8\x00\x1c"
"\x4b\x62\x87\x7c\x15\x38\xda\x70"
"\x3d\xea\xe7\xf2\x40\xba\xae\x79"
"\x8f\x48\xfc\xbf\x45\x53\x2e\x78"
"\xef\x79\xf0\x1b\x49\xf7\xfd\x9c",
.rlen = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x90\x96\x36\xea\x03\x74\x18\x7a"
"\x1d\x46\x42\x2d\x3f\x8c\x54\xc7"
"\x4b\x4a\x73\x89\xa4\x00\x3f\x34"
"\x2c\xb1\xdb\x0c\x44\xe0\xe8\xd2",
.klen = 32,
.iv = "\xa0\x5f\xc5\x52\x47\x13\xc2\x01"
"\x3d\x7c\x6e\x52\x3d\x55\x85\x48",
.assoc = "\xaf",
.alen = 1,
.input = "\x9b\xc5\x3b\x20\x0a\x88\x56\xbe"
"\x69\xdf\xc4\xc4\x02\x46\x3a\xf0",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xcd\xbb\x70\x89\x13\xf0\xc1\x95"
"\x9e\x20\xf4\xbf\x39\xb1\x17\xcd"
"\x76\x0c\x7f\x0d\xa9\xa0\xc1\x4e"
"\xed\xdf\xb9\xe4\x1e\x3f\x87\xa8",
.klen = 32,
.iv = "\xdc\x84\xfe\xf1\x58\x8f\x6b\x1c"
"\xbe\x57\x20\xe3\x37\x7a\x48\x4f",
.assoc = "\xeb\x4d\x8d\x59\x9c\x2e\x15\xa3"
"\xde\x8d\x4d\x07\x36\x43\x78\xd0"
"\x0b\x6d\x84\x4f\x2c\xf0\x82\x5b"
"\x4e\xf6\x29\xd1\x8b\x6f\x56",
.alen = 31,
.input = "\xe0\x6d\xa1\x07\x98\x2f\x40\x2d"
"\x2e\x9a\xd6\x61\x43\xc0\x74\x69",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x0a\xe0\xaa\x29\x24\x6c\x6a\xb1"
"\x1f\xfa\xa6\x50\x33\xd5\xda\xd3"
"\xa0\xce\x8a\x91\xae\x40\x43\x68"
"\xae\x0d\x98\xbd\xf8\x9e\x26\x7f",
.klen = 32,
.iv = "\x19\xa9\x38\x91\x68\x0b\x14\x38"
"\x3f\x31\xd2\x74\x31\x9e\x0a\x55",
.assoc = "\x28\x72\xc7\xf8\xac\xaa\xbe\xbf"
"\x5f\x67\xff\x99\x30\x67\x3b\xd6"
"\x35\x2f\x90\xd3\x31\x90\x04\x74"
"\x0f\x23\x08\xa9\x65\xce\xf6\xea",
.alen = 32,
.input = "\xb9\x57\x13\x3e\x82\x31\x61\x65"
"\x0d\x7f\x6c\x96\x93\x5c\x50\xe2",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x46\x04\xe3\xc8\x34\xe7\x12\xcd"
"\xa0\xd4\x58\xe2\x2d\xf9\x9c\xda"
"\xca\x91\x96\x15\xb4\xe0\xc5\x81"
"\x70\x3a\x77\x95\xd2\xfd\xc5\x55",
.klen = 32,
.iv = "\x55\xcd\x72\x30\x78\x86\xbd\x54"
"\xc0\x0b\x84\x06\x2b\xc2\xcd\x5b",
.assoc = "\x64\x97\x00\x98\xbc\x25\x67\xdb"
"\xe0\x41\xb1\x2a\x2a\x8c\xfe\xdd"
"\x5f\xf2\x9c\x58\x36\x30\x86\x8e"
"\xd1\x51\xe6\x81\x3f\x2d\x95\xc1"
"\x01",
.alen = 33,
.input = "\x81\x96\x34\xde\xbb\x36\xdd\x3e"
"\x4e\x5e\xcb\x44\x21\xb8\x3f\xf1",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x83\x29\x1d\x67\x44\x63\xbb\xe9"
"\x20\xaf\x0a\x73\x27\x1e\x5f\xe0"
"\xf5\x53\xa1\x9a\xb9\x80\x47\x9b"
"\x31\x68\x56\x6e\xac\x5c\x65\x2c",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x92\xf2\xac\xcf\x88\x02\x65\x70"
"\x41\xe5\x36\x97\x25\xe7\x90\x61",
.assoc = "\xa1\xbb\x3a\x37\xcc\xa1\x10\xf7"
"\x61\x1c\x63\xbc\x24\xb0\xc0\xe3"
"\x8a\xb4\xa7\xdc\x3b\xd0\x08\xa8"
"\x92\x7f\xc5\x5a\x19\x8c\x34\x97"
"\x0f\x95\x9b\x18\xe4\x8d\xb4\x24"
"\xb9\x33\x28\x18\xe1\x9d\x14\xe0"
"\x64\xb2\x89\x7d\x78\xa8\x05\x7e"
"\x07\x8c\xfc\x88\x2d\xb8\x53",
.alen = 63,
.input = "\x2e\x99\xb6\x79\x57\x56\x80\x36"
"\x8e\xc4\x1c\x12\x7d\x71\x36\x0c",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xbf\x4e\x57\x07\x54\xdf\x64\x05"
"\xa1\x89\xbc\x04\x21\x42\x22\xe6"
"\x1f\x15\xad\x1e\xbe\x20\xc9\xb4"
"\xf3\x95\x35\x46\x86\xbb\x04\x03",
.klen = 32,
.iv = "\xce\x17\xe5\x6f\x98\x7e\x0e\x8c"
"\xc2\xbf\xe8\x29\x1f\x0b\x52\x68",
.assoc = "\xdd\xe0\x74\xd6\xdc\x1d\xb8\x13"
"\xe2\xf6\x15\x4d\x1e\xd4\x83\xe9"
"\xb4\x76\xb3\x60\x40\x70\x8a\xc1"
"\x53\xac\xa4\x32\xf3\xeb\xd3\x6e"
"\x1e\x42\xa0\x46\x45\x9f\xc7\x22"
"\xd3\x43\xbc\x7e\xa5\x47\x2a\x6f"
"\x91\x19\x70\x1e\xe1\xfe\x25\x49"
"\xd6\x8f\x93\xc7\x28\x3f\x3d\x03",
.alen = 64,
.input = "\x7b\x25\x3d\x47\xd4\xa7\x08\xce"
"\x3b\x89\x40\x36\xba\x6d\x0e\xa2",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.ilen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.result = "",
.rlen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xfc\x72\x90\xa6\x64\x5a\x0d\x21"
"\x22\x63\x6e\x96\x1b\x67\xe4\xec"
"\x49\xd7\xb9\xa2\xc3\xc0\x4b\xce"
"\xb4\xc3\x14\x1e\x61\x1a\xa3\xd9",
.klen = 32,
.iv = "\x0b\x3c\x1f\x0e\xa8\xf9\xb7\xa7"
"\x42\x9a\x9a\xba\x19\x30\x15\x6e",
.assoc = "\x1a",
.alen = 1,
.input = "\xe6\x09\x6f\x95\x9a\x18\xc8\xf6"
"\x17\x75\x81\x16\xdf\x26\xff\x67"
"\x92",
.ilen = 17,
.result = "\x29",
.rlen = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x38\x97\xca\x45\x74\xd6\xb6\x3c"
"\xa3\x3d\x20\x27\x15\x8b\xa7\xf2"
"\x74\x9a\xc4\x27\xc8\x60\xcd\xe8"
"\x75\xf0\xf2\xf7\x3b\x79\x42\xb0",
.klen = 32,
.iv = "\x47\x60\x59\xad\xb8\x75\x60\xc3"
"\xc3\x74\x4c\x4c\x13\x54\xd8\x74",
.assoc = "\x56\x29\xe7\x15\xfc\x14\x0a\x4a"
"\xe4\xaa\x79\x70\x12\x1d\x08\xf6"
"\x09\xfb\xca\x69\x4b\xb0\x8e\xf5"
"\xd6\x07\x62\xe3\xa8\xa9\x12",
.alen = 31,
.input = "\x82\xc0\x56\xf0\xd7\xc4\xc9\xfd"
"\x3c\xd1\x2a\xd4\x15\x86\x9d\xda"
"\xea\x6c\x6f\xa1\x33\xb0\x7a\x01"
"\x57\xe7\xf3\x7b\x73\xe7\x54\x10"
"\xc6\x91\xe2\xc6\xa0\x69\xe7\xe6"
"\x76\xc3\xf5\x3a\x76\xfd\x4a",
.ilen = 47,
.result = "\x66\xf3\x75\x7d\x40\xb3\xb4\xd1"
"\x04\xe1\xa6\x94\x10\xe6\x39\x77"
"\xd3\xac\x4d\x8a\x8c\x58\x6e\xfb"
"\x06\x13\x9a\xd9\x5e\xc0\xfa",
.rlen = 31,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x75\xbc\x04\xe5\x84\x52\x5e\x58"
"\x24\x17\xd2\xb9\x0e\xaf\x6a\xf9"
"\x9e\x5c\xd0\xab\xcd\x00\x4f\x01"
"\x37\x1e\xd1\xcf\x15\xd8\xe2\x86",
.klen = 32,
.iv = "\x84\x85\x92\x4d\xc8\xf1\x08\xdf"
"\x44\x4e\xff\xdd\x0d\x78\x9a\x7a",
.assoc = "\x93\x4e\x21\xb4\x0c\x90\xb3\x66"
"\x65\x84\x2b\x01\x0b\x42\xcb\xfc"
"\x33\xbd\xd6\xed\x50\x50\x10\x0e"
"\x97\x35\x41\xbb\x82\x08\xb1\xf2",
.alen = 32,
.input = "\x01\x47\x8e\x6c\xf6\x64\x89\x3a"
"\x71\xce\xe4\xaa\x45\x70\xe6\x84"
"\x62\x48\x08\x64\x86\x6a\xdf\xec"
"\xb4\xa0\xfb\x34\x03\x0c\x19\xf4"
"\x2b\x7b\x36\x73\xec\x54\xa9\x1e"
"\x30\x85\xdb\xe4\xac\xe9\x2c\xca",
.ilen = 48,
.result = "\xa2\x17\xaf\x1c\x50\x2e\x5d\xed"
"\x85\xbb\x58\x26\x0a\x0b\xfc\x7d"
"\xfe\x6e\x59\x0e\x91\xf8\xf0\x15"
"\xc8\x40\x78\xb1\x38\x1f\x99\xa7",
.rlen = 32,
}, {
.key = "\xb1\xe1\x3e\x84\x94\xcd\x07\x74"
"\xa5\xf2\x84\x4a\x08\xd4\x2c\xff"
"\xc8\x1e\xdb\x2f\xd2\xa0\xd1\x1b"
"\xf8\x4c\xb0\xa8\xef\x37\x81\x5d",
.klen = 32,
.iv = "\xc0\xaa\xcc\xec\xd8\x6c\xb1\xfb"
"\xc5\x28\xb1\x6e\x07\x9d\x5d\x81",
.assoc = "\xd0\x73\x5a\x54\x1d\x0b\x5b\x82"
"\xe5\x5f\xdd\x93\x05\x66\x8e\x02"
"\x5e\x80\xe1\x71\x55\xf0\x92\x28"
"\x59\x62\x20\x94\x5c\x67\x50\xc8"
"\x58",
.alen = 33,
.input = "\x85\xe0\xf8\x0f\x8e\x49\xe3\x60"
"\xcb\x4a\x54\x94\xcf\xf5\x7e\x34"
"\xe9\xf8\x80\x65\x53\xd0\x72\x70"
"\x4f\x7d\x9d\xd1\x15\x6f\xb9\x2c"
"\xfa\xe8\xdd\xac\x2e\xe1\x3f\x67"
"\x63\x0f\x1a\x59\xb7\x89\xdb\xf4"
"\xc3",
.ilen = 49,
.result = "\xdf\x3c\xe9\xbc\x61\xaa\x06\x09"
"\x06\x95\x0a\xb7\x04\x2f\xbe\x84"
"\x28\x30\x64\x92\x96\x98\x72\x2e"
"\x89\x6e\x57\x8a\x13\x7e\x38\x7e"
"\xdb",
.rlen = 33,
}, {
.key = "\xee\x05\x77\x23\xa5\x49\xb0\x90"
"\x26\xcc\x36\xdc\x02\xf8\xef\x05"
"\xf3\xe1\xe7\xb3\xd8\x40\x53\x35"
"\xb9\x79\x8f\x80\xc9\x96\x20\x33",
.klen = 32,
.iv = "\xfd\xce\x06\x8b\xe9\xe8\x5a\x17"
"\x46\x02\x63\x00\x01\xc1\x20\x87",
.assoc = "\x0c\x98\x94\xf3\x2d\x87\x04\x9e"
"\x66\x39\x8f\x24\xff\x8a\x50\x08"
"\x88\x42\xed\xf6\x5a\x90\x14\x42"
"\x1a\x90\xfe\x6c\x36\xc6\xf0\x9f"
"\x66\xa0\xb5\x2d\x2c\xf8\x25\x15"
"\x55\x90\xa2\x7e\x77\x94\x96\x3a"
"\x71\x1c\xf7\x44\xee\xa8\xc3\x42"
"\xe2\xa3\x84\x04\x0b\xe1\xce",
.alen = 63,
.input = "\x00\xe5\x5b\x87\x5c\x20\x22\x8a"
"\xda\x1f\xd3\xff\xbb\xb2\xb0\xf8"
"\xef\xe9\xeb\x9e\x7c\x80\xf4\x2b"
"\x59\xc0\x79\xbc\x17\xa0\x15\x01"
"\xf5\x72\xfb\x5a\xe7\xaf\x07\xe3"
"\x1b\x49\x21\x34\x23\x63\x55\x5e"
"\xee\x4f\x34\x17\xfa\xfe\xa5\x0c"
"\xed\x0b\x23\xea\x9b\xda\x57\x2f"
"\xf6\xa9\xae\x0d\x4e\x40\x96\x45"
"\x7f\xfa\xf0\xbf\xc4\x98\x78",
.ilen = 79,
.result = "\x1b\x61\x23\x5b\x71\x26\xae\x25"
"\x87\x6f\xbc\x49\xfe\x53\x81\x8a"
"\x53\xf2\x70\x17\x9b\x38\xf4\x48"
"\x4b\x9b\x36\x62\xed\xdd\xd8\x54"
"\xea\xcb\xb6\x79\x45\xfc\xaa\x54"
"\x5c\x94\x47\x58\xa7\xff\x9c\x9e"
"\x7c\xb6\xf1\xac\xc8\xfd\x8b\x35"
"\xd5\xa4\x6a\xd4\x09\xc2\x08",
.rlen = 63,
}, {
.key = "\x2a\x2a\xb1\xc3\xb5\xc5\x59\xac"
"\xa7\xa6\xe8\x6d\xfc\x1d\xb2\x0b"
"\x1d\xa3\xf3\x38\xdd\xe0\xd5\x4e"
"\x7b\xa7\x6e\x58\xa3\xf5\xbf\x0a",
.klen = 32,
.iv = "\x39\xf3\x3f\x2b\xf9\x64\x03\x33"
"\xc7\xdd\x15\x91\xfb\xe6\xe2\x8d",
.assoc = "\x49\xbc\xce\x92\x3d\x02\xad\xba"
"\xe7\x13\x41\xb6\xf9\xaf\x13\x0f"
"\xb2\x04\xf8\x7a\x5f\x30\x96\x5b"
"\xdc\xbd\xdd\x44\x10\x25\x8f\x75"
"\x75\x4d\xb9\x5b\x8e\x0a\x38\x13"
"\x6f\x9f\x36\xe4\x3a\x3e\xac\xc9"
"\x9d\x83\xde\xe5\x57\xfd\xe3\x0e"
"\xb1\xa7\x1b\x44\x05\x67\xb7\x37",
.alen = 64,
.input = "\x28\xdd\xb9\x4a\x12\xc7\x0a\xe1"
"\x58\x06\x1a\x9b\x8c\x67\xdf\xeb"
"\x35\x35\x60\x9d\x06\x40\x65\xc1"
"\x93\xe8\xb3\x82\x50\x29\xdd\xb5"
"\x2b\xcb\xde\x18\x78\x6b\x42\xbe"
"\x6d\x24\xd0\xb2\x7d\xd7\x08\x8f"
"\x4a\x18\x98\xad\x8c\xf2\x97\xb4"
"\xf4\x77\xe4\xbf\x41\x3b\xc4\x06"
"\xce\x9e\x34\x81\xf0\x89\x11\x13"
"\x02\x65\xa1\x7c\xdf\x07\x33\x06",
.ilen = 80,
.result = "\x58\x85\x5c\xfa\x81\xa1\x57\x40"
"\x08\x4a\x6e\xda\xf8\x78\x44\x90"
"\x7d\xb5\x7b\x9b\xa1\xd8\x76\x62"
"\x0c\xc9\x15\x3b\xc7\x3c\x77\x2b"
"\xf8\x78\xba\xa7\xa6\x0e\xbd\x52"
"\x76\xa3\xdc\xbe\x6b\xa8\xb1\x2d"
"\xa9\x1d\xd8\x4e\x31\x53\xab\x00"
"\xa5\xa7\x01\x13\x04\x49\xf2\x04",
.rlen = 64,
}, {
.key = "\x67\x4f\xeb\x62\xc5\x40\x01\xc7"
"\x28\x80\x9a\xfe\xf6\x41\x74\x12"
"\x48\x65\xfe\xbc\xe2\x80\x57\x68"
"\x3c\xd4\x4d\x31\x7d\x54\x5f\xe1",
.klen = 32,
.iv = "\x76\x18\x79\xca\x09\xdf\xac\x4e"
"\x48\xb7\xc7\x23\xf5\x0a\xa5\x93",
.assoc = "\x85\xe1\x08\x32\x4d\x7e\x56\xd5"
"\x68\xed\xf3\x47\xf3\xd3\xd6\x15"
"\xdd\xc7\x04\xfe\x64\xd0\x18\x75"
"\x9d\xeb\xbc\x1d\xea\x84\x2e\x4c"
"\x83\xf9\xbe\x8a\xef\x1c\x4b\x10"
"\x89\xaf\xcb\x4b\xfe\xe7\xc1\x58"
"\xca\xea\xc6\x87\xc0\x53\x03\xd9"
"\x80\xaa\xb2\x83\xff\xee\xa1\x6a"
"\x04",
.alen = 65,
.input = "\x85\x39\x69\x35\xfb\xf9\xb0\xa6"
"\x85\x43\x88\xd0\xd7\x78\x60\x19"
"\x3e\x1f\xb1\xa4\xd6\xc5\x96\xec"
"\xf7\x84\x85\xc7\x27\x0f\x74\x57"
"\x28\x9e\xdd\x90\x3c\x43\x12\xc5"
"\x51\x3d\x39\x8f\xa5\xf4\xe0\x0b"
"\x57\x04\xf1\x6d\xfe\x9b\x84\x27"
"\xe8\xeb\x4d\xda\x02\x0a\xc5\x49"
"\x1a\x55\x5e\x50\x56\x4d\x94\xda"
"\x20\xf8\x12\x54\x50\xb3\x11\xda"
"\xed\x44\x27\x67\xd5\xd1\x8b\x4b"
"\x38\x67\x56\x65\x59\xda\xe6\x97"
"\x81\xae\x2f\x92\x3b\xae\x22\x1c"
"\x91\x59\x38\x18\x00\xe8\xba\x92"
"\x04\x19\x56\xdf\xb0\x82\xeb\x6f"
"\x2e\xdb\x54\x3c\x4b\xbb\x60\x90"
"\x4c\x50\x10\x62\xba\x7a\xb1\x68"
"\x37\xd7\x87\x4e\xe4\x66\x09\x1f"
"\xa5",
.ilen = 145,
.result = "\x94\xaa\x96\x9a\x91\x1d\x00\x5c"
"\x88\x24\x20\x6b\xf2\x9c\x06\x96"
"\xa7\x77\x87\x1f\xa6\x78\xf8\x7b"
"\xcd\xf6\xf4\x13\xa1\x9b\x16\x02"
"\x07\x24\xbf\xd5\x08\x20\xd0\x4f"
"\x90\xb3\x70\x24\x2f\x51\xc7\xbb"
"\xd6\x84\xc0\xef\x9a\xa8\xca\xcc"
"\x74\xab\x97\x53\xfe\xd0\xdb\x37"
"\x37\x6a\x0e\x9f\x3f\xa3\x2a\xe3"
"\x1b\x34\x6d\x51\x72\x2b\x17\xe7"
"\x4d\xaa\x2c\x18\xda\xa3\x33\x89"
"\x2a\x9f\xf4\xd2\xed\x76\x3d\x3f"
"\x3c\x15\x9d\x8e\x4f\x3c\x27\xb0"
"\x42\x3f\x2f\x8a\xd4\xc2\x10\xb2"
"\x27\x7f\xe3\x34\x80\x02\x49\x4b"
"\x07\x68\x22\x2a\x88\x25\x53\xb2"
"\x2f",
.rlen = 129,
}, {
.key = "\xa3\x73\x24\x01\xd5\xbc\xaa\xe3"
"\xa9\x5a\x4c\x90\xf0\x65\x37\x18"
"\x72\x28\x0a\x40\xe7\x20\xd9\x82"
"\xfe\x02\x2b\x09\x57\xb3\xfe\xb7",
.klen = 32,
.iv = "\xb3\x3d\xb3\x69\x19\x5b\x54\x6a"
"\xc9\x91\x79\xb4\xef\x2e\x68\x99",
.assoc = "\xc2\x06\x41\xd1\x5d\xfa\xff\xf1"
"\xe9\xc7\xa5\xd9\xed\xf8\x98\x1b"
"\x07\x89\x10\x82\x6a\x70\x9a\x8f"
"\x5e\x19\x9b\xf5\xc5\xe3\xcd\x22"
"\x92\xa5\xc2\xb8\x51\x2e\x5e\x0e"
"\xa4\xbe\x5f\xb1\xc1\x90\xd7\xe7"
"\xf7\x52\xae\x28\x29\xa8\x22\xa4"
"\x4f\xae\x48\xc2\xfa\x75\x8b\x9e"
"\xce\x83\x2a\x88\x07\x55\xbb\x89"
"\xf6\xdf\xac\xdf\x83\x08\xbf\x7d"
"\xac\x30\x8b\x8e\x02\xac\x00\xf1"
"\x30\x46\xe1\xbc\x75\xbf\x49\xbb"
"\x26\x4e\x29\xf0\x2f\x21\xc6\x13"
"\x92\xd9\x3d\x11\xe4\x10\x00\x8e"
"\xd4\xd4\x58\x65\xa6\x2b\xe3\x25"
"\xb1\x8f\x15\x93\xe7\x71\xb9\x2c"
"\x4b",
.alen = 129,
.input = "\x7d\xde\x53\x22\xe4\x23\x3b\x30"
"\x78\xde\x35\x90\x7a\xd9\x0b\x93"
"\xf6\x0e\x0b\xed\x40\xee\x10\x9c"
"\x96\x3a\xd3\x34\xb2\xd0\x67\xcf"
"\x63\x7f\x2d\x0c\xcf\x96\xec\x64"
"\x1a\x87\xcc\x7d\x2c\x5e\x81\x4b"
"\xd2\x8f\x4c\x7c\x00\xb1\xb4\xe0"
"\x87\x4d\xb1\xbc\xd8\x78\x2c\x17"
"\xf2\x3b\xd8\x28\x40\xe2\x76\xf6"
"\x20\x13\x83\x46\xaf\xff\xe3\x0f"
"\x72",
.ilen = 81,
.result = "\xd1\xcf\xd0\x39\xa1\x99\xa9\x78"
"\x09\xfe\xd2\xfd\xec\xc1\xc9\x9d"
"\xd2\x39\x93\xa3\xab\x18\x7a\x95"
"\x8f\x24\xd3\xeb\x7b\xfa\xb5\xd8"
"\x15\xd1\xc3\x04\x69\x32\xe3\x4d"
"\xaa\xc2\x04\x8b\xf2\xfa\xdc\x4a"
"\x02\xeb\xa8\x90\x03\xfd\xea\x97"
"\x43\xaf\x2e\x92\xf8\x57\xc5\x6a"
"\x00",
.rlen = 65,
}, {
.key = "\xe0\x98\x5e\xa1\xe5\x38\x53\xff"
"\x2a\x35\xfe\x21\xea\x8a\xfa\x1e"
"\x9c\xea\x15\xc5\xec\xc0\x5b\x9b"
"\xbf\x2f\x0a\xe1\x32\x12\x9d\x8e",
.klen = 32,
.iv = "\xef\x61\xed\x08\x29\xd7\xfd\x86"
"\x4a\x6b\x2b\x46\xe9\x53\x2a\xa0",
.assoc = "\xfe\x2a\x7b\x70\x6d\x75\xa7\x0d"
"\x6a\xa2\x57\x6a\xe7\x1c\x5b\x21"
"\x31\x4b\x1b\x07\x6f\x10\x1c\xa8"
"\x20\x46\x7a\xce\x9f\x42\x6d\xf9",
.alen = 32,
.input = "\x5a\xcd\x8c\x57\xf2\x6a\xb6\xbe"
"\x53\xc7\xaa\x9a\x60\x74\x9c\xc4"
"\xa2\xc2\xd0\x6d\xe1\x03\x63\xdc"
"\xbb\x51\x7e\x9c\x89\x73\xde\x4e"
"\x24\xf8\x52\x7c\x15\x41\x0e\xba"
"\x69\x0e\x36\x5f\x2f\x22\x8c",
.ilen = 47,
.result = "\x0d\xf4\x09\xd8\xb1\x14\x51\x94"
"\x8a\xd8\x84\x8e\xe6\xe5\x8c\xa3"
"\xfc\xfc\x9e\x28\xb0\xb8\xfc\xaf"
"\x50\x52\xb1\xc4\x55\x59\x55\xaf",
.rlen = 32,
}, {
.key = "\x1c\xbd\x98\x40\xf5\xb3\xfc\x1b"
"\xaa\x0f\xb0\xb3\xe4\xae\xbc\x24"
"\xc7\xac\x21\x49\xf1\x60\xdd\xb5"
"\x80\x5d\xe9\xba\x0c\x71\x3c\x64",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\x2c\x86\x26\xa8\x39\x52\xa6\xa2"
"\xcb\x45\xdd\xd7\xe3\x77\xed\xa6",
.assoc = "\x3b\x4f\xb5\x10\x7d\xf1\x50\x29"
"\xeb\x7c\x0a\xfb\xe1\x40\x1e\x27"
"\x5c\x0d\x27\x8b\x74\xb0\x9e\xc2"
"\xe1\x74\x59\xa6\x79\xa1\x0c\xd0",
.alen = 32,
.input = "\x47\xd6\xce\x78\xd6\xbf\x4a\x51"
"\xb8\xda\x92\x3c\xfd\xda\xac\x8e"
"\x8d\x88\xd7\x4d\x90\xe5\xeb\xa1"
"\xab\xd6\x7c\x76\xad\xea\x7d\x76"
"\x53\xee\xb0\xcd\xd0\x02\xbb\x70"
"\x5b\x6f\x7b\xe2\x8c\xe8",
.ilen = 46,
.result = "\x4a\x18\x43\x77\xc1\x90\xfa\xb0"
"\x0b\xb2\x36\x20\xe0\x09\x4e\xa9"
"\x26\xbe\xaa\xac\xb5\x58\x7e\xc8"
"\x11\x7f\x90\x9c\x2f\xb8\xf4\x85",
.rlen = 32,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x59\xe1\xd2\xdf\x05\x2f\xa4\x37"
"\x2b\xe9\x63\x44\xde\xd3\x7f\x2b"
"\xf1\x6f\x2d\xcd\xf6\x00\x5f\xcf"
"\x42\x8a\xc8\x92\xe6\xd0\xdc\x3b",
.klen = 32,
.iv = "\x68\xab\x60\x47\x49\xce\x4f\xbe"
"\x4c\x20\x8f\x68\xdd\x9c\xb0\xac",
.assoc = "\x77\x74\xee\xaf\x8d\x6d\xf9\x45"
"\x6c\x56\xbc\x8d\xdb\x65\xe0\x2e"
"\x86\xd0\x32\x0f\x79\x50\x20\xdb"
"\xa2\xa1\x37\x7e\x53\x00\xab\xa6",
.alen = 32,
.input = "\x9f\xa9\x2b\xa4\x8f\x00\x05\x2b"
"\xe7\x68\x81\x51\xbb\xfb\xdf\x60"
"\xbb\xac\xe8\xc1\xdc\x68\xae\x68"
"\x3a\xcd\x7a\x06\x49\xfe\x80\x11"
"\xe6\x61\x99\xe2\xdd\xbe\x2c\xbf",
.ilen = 40,
.result = "\x86\x3d\x7d\x17\xd1\x0c\xa3\xcc"
"\x8c\x8d\xe8\xb1\xda\x2e\x11\xaf"
"\x51\x80\xb5\x30\xba\xf8\x00\xe2"
"\xd3\xad\x6f\x75\x09\x18\x93\x5c",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 32,
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x96\x06\x0b\x7f\x15\xab\x4d\x53"
"\xac\xc3\x15\xd6\xd8\xf7\x42\x31"
"\x1b\x31\x38\x51\xfc\xa0\xe1\xe8"
"\x03\xb8\xa7\x6b\xc0\x2f\x7b\x11",
.klen = 32,
.iv = "\xa5\xcf\x9a\xe6\x59\x4a\xf7\xd9"
"\xcd\xfa\x41\xfa\xd7\xc0\x72\xb2",
.assoc = "\xb4\x99\x28\x4e\x9d\xe8\xa2\x60"
"\xed\x30\x6e\x1e\xd5\x89\xa3\x34"
"\xb1\x92\x3e\x93\x7e\xf0\xa2\xf5"
"\x64\xcf\x16\x57\x2d\x5f\x4a\x7d",
.alen = 32,
.input = "\xe2\x34\xfa\x25\xfd\xfb\x89\x5e"
"\x5b\x4e\x0b\x15\x6e\x39\xfb\x0c"
"\x73\xc7\xd9\x6b\xbe\xce\x9b\x70"
"\xc7\x4f\x96\x16\x03\xfc\xea\xfb"
"\x56",
.ilen = 33,
.result = "\xc3\x62\xb7\xb6\xe2\x87\x4c\xe7"
"\x0d\x67\x9a\x43\xd4\x52\xd4\xb5"
"\x7b\x43\xc1\xb5\xbf\x98\x82\xfc"
"\x94\xda\x4e\x4d\xe4\x77\x32\x32",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 32,
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* All key wrapping test vectors taken from
* http://csrc.nist.gov/groups/STM/cavp/documents/mac/kwtestvectors.zip
*
* Note: as documented in keywrap.c, the ivout for encryption is the first
* semiblock of the ciphertext from the test vector. For decryption, iv is
* the first semiblock of the ciphertext.
*/
static const struct cipher_testvec aes_kw_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x75\x75\xda\x3a\x93\x60\x7c\xc2"
"\xbf\xd8\xce\xc7\xaa\xdf\xd9\xa6",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x42\x13\x6d\x3c\x38\x4a\x3e\xea"
"\xc9\x5a\x06\x6f\xd2\x8f\xed\x3f",
.ctext = "\xf6\x85\x94\x81\x6f\x64\xca\xa3"
"\xf5\x6f\xab\xea\x25\x48\xf5\xfb",
.len = 16,
.iv = "\x03\x1f\x6b\xd7\xe6\x1e\x64\x3d",
.generates_iv = true,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x80\xaa\x99\x73\x27\xa4\x80\x6b"
"\x6a\x7a\x41\xa5\x2b\x86\xc3\x71"
"\x03\x86\xf9\x32\x78\x6e\xf7\x96"
"\x76\xfa\xfb\x90\xb8\x26\x3c\x5f",
.klen = 32,
.ptext = "\x0a\x25\x6b\xa7\x5c\xfa\x03\xaa"
"\xa0\x2b\xa9\x42\x03\xf1\x5b\xaa",
.ctext = "\xd3\x3d\x3d\x97\x7b\xf0\xa9\x15"
"\x59\xf9\x9c\x8a\xcd\x29\x3d\x43",
.len = 16,
.iv = "\x42\x3c\x96\x0d\x8a\x2a\xc4\xc1",
.generates_iv = true,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* ANSI X9.31 Continuous Pseudo-Random Number Generator (AES mode)
* test vectors, taken from Appendix B.2.9 and B.2.10:
* http://csrc.nist.gov/groups/STM/cavp/documents/rng/RNGVS.pdf
* Only AES-128 is supported at this time.
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cprng_testvec ansi_cprng_aes_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xb1\x66\x6d\x13\x60\x72\x42"
"\xed\x06\x1c\xab\xb8\xd4\x62\x02",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\xe6\xb3\xbe\x78\x2a\x23\xfa\x62"
"\xd7\x1d\x4a\xfb\xb0\xe9\x22\xf9",
.dtlen = 16,
.v = "\x80\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.vlen = 16,
.result = "\x59\x53\x1e\xd1\x3b\xb0\xc0\x55"
"\x84\x79\x66\x85\xc1\x2f\x76\x41",
.rlen = 16,
.loops = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xb1\x66\x6d\x13\x60\x72\x42"
"\xed\x06\x1c\xab\xb8\xd4\x62\x02",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\xe6\xb3\xbe\x78\x2a\x23\xfa\x62"
"\xd7\x1d\x4a\xfb\xb0\xe9\x22\xfa",
.dtlen = 16,
.v = "\xc0\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.vlen = 16,
.result = "\x7c\x22\x2c\xf4\xca\x8f\xa2\x4c"
"\x1c\x9c\xb6\x41\xa9\xf3\x22\x0d",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.loops = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xb1\x66\x6d\x13\x60\x72\x42"
"\xed\x06\x1c\xab\xb8\xd4\x62\x02",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\xe6\xb3\xbe\x78\x2a\x23\xfa\x62"
"\xd7\x1d\x4a\xfb\xb0\xe9\x22\xfb",
.dtlen = 16,
.v = "\xe0\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.vlen = 16,
.result = "\x8a\xaa\x00\x39\x66\x67\x5b\xe5"
"\x29\x14\x28\x81\xa9\x4d\x4e\xc7",
.rlen = 16,
.loops = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xb1\x66\x6d\x13\x60\x72\x42"
"\xed\x06\x1c\xab\xb8\xd4\x62\x02",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\xe6\xb3\xbe\x78\x2a\x23\xfa\x62"
"\xd7\x1d\x4a\xfb\xb0\xe9\x22\xfc",
.dtlen = 16,
.v = "\xf0\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.vlen = 16,
.result = "\x88\xdd\xa4\x56\x30\x24\x23\xe5"
"\xf6\x9d\xa5\x7e\x7b\x95\xc7\x3a",
.rlen = 16,
.loops = 1,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\xf3\xb1\x66\x6d\x13\x60\x72\x42"
"\xed\x06\x1c\xab\xb8\xd4\x62\x02",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\xe6\xb3\xbe\x78\x2a\x23\xfa\x62"
"\xd7\x1d\x4a\xfb\xb0\xe9\x22\xfd",
.dtlen = 16,
.v = "\xf8\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.vlen = 16,
.result = "\x05\x25\x92\x46\x61\x79\xd2\xcb"
"\x78\xc4\x0b\x14\x0a\x5a\x9a\xc8",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.rlen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.loops = 1,
}, { /* Monte Carlo Test */
.key = "\x9f\x5b\x51\x20\x0b\xf3\x34\xb5"
"\xd8\x2b\xe8\xc3\x72\x55\xc8\x48",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.dt = "\x63\x76\xbb\xe5\x29\x02\xba\x3b"
"\x67\xc9\x25\xfa\x70\x1f\x11\xac",
.dtlen = 16,
.v = "\x57\x2c\x8e\x76\x87\x26\x47\x97"
"\x7e\x74\xfb\xdd\xc4\x95\x01\xd1",
.vlen = 16,
.result = "\x48\xe9\xbd\x0d\x06\xee\x18\xfb"
"\xe4\x57\x90\xd5\xc3\xfc\x9b\x73",
.rlen = 16,
.loops = 10000,
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
/*
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
* SP800-90A DRBG Test vectors from
* http://csrc.nist.gov/groups/STM/cavp/documents/drbg/drbgtestvectors.zip
*
* Test vectors for DRBG with prediction resistance. All types of DRBGs
* (Hash, HMAC, CTR) are tested with all permutations of use cases (w/ and
* w/o personalization string, w/ and w/o additional input string).
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct drbg_testvec drbg_pr_sha256_tv_template[] = {
{
.entropy = (unsigned char *)
"\x72\x88\x4c\xcd\x6c\x85\x57\x70\xf7\x0b\x8b\x86"
"\xc1\xeb\xd2\x4e\x36\x14\xab\x18\xc4\x9c\xc9\xcf"
"\x1a\xe8\xf7\x7b\x02\x49\x73\xd7\xf1\x42\x7d\xc6"
"\x3f\x29\x2d\xec\xd3\x66\x51\x3f\x1d\x8d\x5b\x4e",
.entropylen = 48,
.entpra = (unsigned char *)
"\x38\x9c\x91\xfa\xc2\xa3\x46\x89\x56\x08\x3f\x62"
"\x73\xd5\x22\xa9\x29\x63\x3a\x1d\xe5\x5d\x5e\x4f"
"\x67\xb0\x67\x7a\x5e\x9e\x0c\x62",
.entprb = (unsigned char *)
"\xb2\x8f\x36\xb2\xf6\x8d\x39\x13\xfa\x6c\x66\xcf"
"\x62\x8a\x7e\x8c\x12\x33\x71\x9c\x69\xe4\xa5\xf0"
"\x8c\xee\xeb\x9c\xf5\x31\x98\x31",
.entprlen = 32,
.expected = (unsigned char *)
"\x52\x7b\xa3\xad\x71\x77\xa4\x49\x42\x04\x61\xc7"
"\xf0\xaf\xa5\xfd\xd3\xb3\x0d\x6a\x61\xba\x35\x49"
"\xbb\xaa\xaf\xe4\x25\x7d\xb5\x48\xaf\x5c\x18\x3d"
"\x33\x8d\x9d\x45\xdf\x98\xd5\x94\xa8\xda\x92\xfe"
"\xc4\x3c\x94\x2a\xcf\x7f\x7b\xf2\xeb\x28\xa9\xf1"
"\xe0\x86\x30\xa8\xfe\xf2\x48\x90\x91\x0c\x75\xb5"
"\x3c\x00\xf0\x4d\x09\x4f\x40\xa7\xa2\x8c\x52\xdf"
"\x52\xef\x17\xbf\x3d\xd1\xa2\x31\xb4\xb8\xdc\xe6"
"\x5b\x0d\x1f\x78\x36\xb4\xe6\x4b\xa7\x11\x25\xd5"
"\x94\xc6\x97\x36\xab\xf0\xe5\x31\x28\x6a\xbb\xce"
"\x30\x81\xa6\x8f\x27\x14\xf8\x1c",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x5d\xf2\x14\xbc\xf6\xb5\x4e\x0b\xf0\x0d\x6f\x2d"
"\xe2\x01\x66\x7b\xd0\xa4\x73\xa4\x21\xdd\xb0\xc0"
"\x51\x79\x09\xf4\xea\xa9\x08\xfa\xa6\x67\xe0\xe1"
"\xd1\x88\xa8\xad\xee\x69\x74\xb3\x55\x06\x9b\xf6",
.entropylen = 48,
.entpra = (unsigned char *)
"\xef\x48\x06\xa2\xc2\x45\xf1\x44\xfa\x34\x2c\xeb"
"\x8d\x78\x3c\x09\x8f\x34\x72\x20\xf2\xe7\xfd\x13"
"\x76\x0a\xf6\xdc\x3c\xf5\xc0\x15",
.entprb = (unsigned char *)
"\x4b\xbe\xe5\x24\xed\x6a\x2d\x0c\xdb\x73\x5e\x09"
"\xf9\xad\x67\x7c\x51\x47\x8b\x6b\x30\x2a\xc6\xde"
"\x76\xaa\x55\x04\x8b\x0a\x72\x95",
.entprlen = 32,
.expected = (unsigned char *)
"\x3b\x14\x71\x99\xa1\xda\xa0\x42\xe6\xc8\x85\x32"
"\x70\x20\x32\x53\x9a\xbe\xd1\x1e\x15\xef\xfb\x4c"
"\x25\x6e\x19\x3a\xf0\xb9\xcb\xde\xf0\x3b\xc6\x18"
"\x4d\x85\x5a\x9b\xf1\xe3\xc2\x23\x03\x93\x08\xdb"
"\xa7\x07\x4b\x33\x78\x40\x4d\xeb\x24\xf5\x6e\x81"
"\x4a\x1b\x6e\xa3\x94\x52\x43\xb0\xaf\x2e\x21\xf4"
"\x42\x46\x8e\x90\xed\x34\x21\x75\xea\xda\x67\xb6"
"\xe4\xf6\xff\xc6\x31\x6c\x9a\x5a\xdb\xb3\x97\x13"
"\x09\xd3\x20\x98\x33\x2d\x6d\xd7\xb5\x6a\xa8\xa9"
"\x9a\x5b\xd6\x87\x52\xa1\x89\x2b\x4b\x9c\x64\x60"
"\x50\x47\xa3\x63\x81\x16\xaf\x19",
.expectedlen = 128,
.addtla = (unsigned char *)
"\xbe\x13\xdb\x2a\xe9\xa8\xfe\x09\x97\xe1\xce\x5d"
"\xe8\xbb\xc0\x7c\x4f\xcb\x62\x19\x3f\x0f\xd2\xad"
"\xa9\xd0\x1d\x59\x02\xc4\xff\x70",
.addtlb = (unsigned char *)
"\x6f\x96\x13\xe2\xa7\xf5\x6c\xfe\xdf\x66\xe3\x31"
"\x63\x76\xbf\x20\x27\x06\x49\xf1\xf3\x01\x77\x41"
"\x9f\xeb\xe4\x38\xfe\x67\x00\xcd",
.addtllen = 32,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\xc6\x1c\xaf\x83\xa2\x56\x38\xf9\xb0\xbc\xd9\x85"
"\xf5\x2e\xc4\x46\x9c\xe1\xb9\x40\x98\x70\x10\x72"
"\xd7\x7d\x15\x85\xa1\x83\x5a\x97\xdf\xc8\xa8\xe8"
"\x03\x4c\xcb\x70\x35\x8b\x90\x94\x46\x8a\x6e\xa1",
.entropylen = 48,
.entpra = (unsigned char *)
"\xc9\x05\xa4\xcf\x28\x80\x4b\x93\x0f\x8b\xc6\xf9"
"\x09\x41\x58\x74\xe9\xec\x28\xc7\x53\x0a\x73\x60"
"\xba\x0a\xde\x57\x5b\x4b\x9f\x29",
.entprb = (unsigned char *)
"\x4f\x31\xd2\xeb\xac\xfa\xa8\xe2\x01\x7d\xf3\xbd"
"\x42\xbd\x20\xa0\x30\x65\x74\xd5\x5d\xd2\xad\xa4"
"\xa9\xeb\x1f\x4d\xf6\xfd\xb8\x26",
.entprlen = 32,
.expected = (unsigned char *)
"\xf6\x13\x05\xcb\x83\x60\x16\x42\x49\x1d\xc6\x25"
"\x3b\x8c\x31\xa3\xbe\x8b\xbd\x1c\xe2\xec\x1d\xde"
"\xbb\xbf\xa1\xac\xa8\x9f\x50\xce\x69\xce\xef\xd5"
"\xd6\xf2\xef\x6a\xf7\x81\x38\xdf\xbc\xa7\x5a\xb9"
"\xb2\x42\x65\xab\xe4\x86\x8d\x2d\x9d\x59\x99\x2c"
"\x5a\x0d\x71\x55\x98\xa4\x45\xc2\x8d\xdb\x05\x5e"
"\x50\x21\xf7\xcd\xe8\x98\x43\xce\x57\x74\x63\x4c"
"\xf3\xb1\xa5\x14\x1e\x9e\x01\xeb\x54\xd9\x56\xae"
"\xbd\xb6\x6f\x1a\x47\x6b\x3b\x44\xe4\xa2\xe9\x3c"
"\x6c\x83\x12\x30\xb8\x78\x7f\x8e\x54\x82\xd4\xfe"
"\x90\x35\x0d\x4c\x4d\x85\xe7\x13",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\xa5\xbf\xac\x4f\x71\xa1\xbb\x67\x94\xc6\x50\xc7"
"\x2a\x45\x9e\x10\xa8\xed\xf7\x52\x4f\xfe\x21\x90"
"\xa4\x1b\xe1\xe2\x53\xcc\x61\x47",
.perslen = 32,
}, {
.entropy = (unsigned char *)
"\xb6\xc1\x8d\xdf\x99\x54\xbe\x95\x10\x48\xd9\xf6"
"\xd7\x48\xa8\x73\x2d\x74\xde\x1e\xde\x57\x7e\xf4"
"\x7b\x7b\x64\xef\x88\x7a\xa8\x10\x4b\xe1\xc1\x87"
"\xbb\x0b\xe1\x39\x39\x50\xaf\x68\x9c\xa2\xbf\x5e",
.entropylen = 48,
.entpra = (unsigned char *)
"\xdc\x81\x0a\x01\x58\xa7\x2e\xce\xee\x48\x8c\x7c"
"\x77\x9e\x3c\xf1\x17\x24\x7a\xbb\xab\x9f\xca\x12"
"\x19\xaf\x97\x2d\x5f\xf9\xff\xfc",
.entprb = (unsigned char *)
"\xaf\xfc\x4f\x98\x8b\x93\x95\xc1\xb5\x8b\x7f\x73"
"\x6d\xa6\xbe\x6d\x33\xeb\x2c\x82\xb1\xaf\xc1\xb6"
"\xb6\x05\xe2\x44\xaa\xfd\xe7\xdb",
.entprlen = 32,
.expected = (unsigned char *)
"\x51\x79\xde\x1c\x0f\x58\xf3\xf4\xc9\x57\x2e\x31"
"\xa7\x09\xa1\x53\x64\x63\xa2\xc5\x1d\x84\x88\x65"
"\x01\x1b\xc6\x16\x3c\x49\x5b\x42\x8e\x53\xf5\x18"
"\xad\x94\x12\x0d\x4f\x55\xcc\x45\x5c\x98\x0f\x42"
"\x28\x2f\x47\x11\xf9\xc4\x01\x97\x6b\xa0\x94\x50"
"\xa9\xd1\x5e\x06\x54\x3f\xdf\xbb\xc4\x98\xee\x8b"
"\xba\xa9\xfa\x49\xee\x1d\xdc\xfb\x50\xf6\x51\x9f"
"\x6c\x4a\x9a\x6f\x63\xa2\x7d\xad\xaf\x3a\x24\xa0"
"\xd9\x9f\x07\xeb\x15\xee\x26\xe0\xd5\x63\x39\xda"
"\x3c\x59\xd6\x33\x6c\x02\xe8\x05\x71\x46\x68\x44"
"\x63\x4a\x68\x72\xe9\xf5\x55\xfe",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x15\x20\x2f\xf6\x98\x28\x63\xa2\xc4\x4e\xbb\x6c"
"\xb2\x25\x92\x61\x79\xc9\x22\xc4\x61\x54\x96\xff"
"\x4a\x85\xca\x80\xfe\x0d\x1c\xd0",
.addtlb = (unsigned char *)
"\xde\x29\x8e\x03\x42\x61\xa3\x28\x5e\xc8\x80\xc2"
"\x6d\xbf\xad\x13\xe1\x8d\x2a\xc7\xe8\xc7\x18\x89"
"\x42\x58\x9e\xd6\xcc\xad\x7b\x1e",
.addtllen = 32,
.pers = (unsigned char *)
"\x84\xc3\x73\x9e\xce\xb3\xbc\x89\xf7\x62\xb3\xe1"
"\xd7\x48\x45\x8a\xa9\xcc\xe9\xed\xd5\x81\x84\x52"
"\x82\x4c\xdc\x19\xb8\xf8\x92\x5c",
.perslen = 32,
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct drbg_testvec drbg_pr_hmac_sha256_tv_template[] = {
{
.entropy = (unsigned char *)
"\x99\x69\xe5\x4b\x47\x03\xff\x31\x78\x5b\x87\x9a"
"\x7e\x5c\x0e\xae\x0d\x3e\x30\x95\x59\xe9\xfe\x96"
"\xb0\x67\x6d\x49\xd5\x91\xea\x4d\x07\xd2\x0d\x46"
"\xd0\x64\x75\x7d\x30\x23\xca\xc2\x37\x61\x27\xab",
.entropylen = 48,
.entpra = (unsigned char *)
"\xc6\x0f\x29\x99\x10\x0f\x73\x8c\x10\xf7\x47\x92"
"\x67\x6a\x3f\xc4\xa2\x62\xd1\x37\x21\x79\x80\x46"
"\xe2\x9a\x29\x51\x81\x56\x9f\x54",
.entprb = (unsigned char *)
"\xc1\x1d\x45\x24\xc9\x07\x1b\xd3\x09\x60\x15\xfc"
"\xf7\xbc\x24\xa6\x07\xf2\x2f\xa0\x65\xc9\x37\x65"
"\x8a\x2a\x77\xa8\x69\x90\x89\xf4",
.entprlen = 32,
.expected = (unsigned char *)
"\xab\xc0\x15\x85\x60\x94\x80\x3a\x93\x8d\xff\xd2"
"\x0d\xa9\x48\x43\x87\x0e\xf9\x35\xb8\x2c\xfe\xc1"
"\x77\x06\xb8\xf5\x51\xb8\x38\x50\x44\x23\x5d\xd4"
"\x4b\x59\x9f\x94\xb3\x9b\xe7\x8d\xd4\x76\xe0\xcf"
"\x11\x30\x9c\x99\x5a\x73\x34\xe0\xa7\x8b\x37\xbc"
"\x95\x86\x23\x50\x86\xfa\x3b\x63\x7b\xa9\x1c\xf8"
"\xfb\x65\xef\xa2\x2a\x58\x9c\x13\x75\x31\xaa\x7b"
"\x2d\x4e\x26\x07\xaa\xc2\x72\x92\xb0\x1c\x69\x8e"
"\x6e\x01\xae\x67\x9e\xb8\x7c\x01\xa8\x9c\x74\x22"
"\xd4\x37\x2d\x6d\x75\x4a\xba\xbb\x4b\xf8\x96\xfc"
"\xb1\xcd\x09\xd6\x92\xd0\x28\x3f",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\xb9\x1f\xe9\xef\xdd\x9b\x7d\x20\xb6\xec\xe0\x2f"
"\xdb\x76\x24\xce\x41\xc8\x3a\x4a\x12\x7f\x3e\x2f"
"\xae\x05\x99\xea\xb5\x06\x71\x0d\x0c\x4c\xb4\x05"
"\x26\xc6\xbd\xf5\x7f\x2a\x3d\xf2\xb5\x49\x7b\xda",
.entropylen = 48,
.entpra = (unsigned char *)
"\xef\x67\x50\x9c\xa7\x7d\xdf\xb7\x2d\x81\x01\xa4"
"\x62\x81\x6a\x69\x5b\xb3\x37\x45\xa7\x34\x8e\x26"
"\x46\xd9\x26\xa2\x19\xd4\x94\x43",
.entprb = (unsigned char *)
"\x97\x75\x53\x53\xba\xb4\xa6\xb2\x91\x60\x71\x79"
"\xd1\x6b\x4a\x24\x9a\x34\x66\xcc\x33\xab\x07\x98"
"\x51\x78\x72\xb2\x79\xfd\x2c\xff",
.entprlen = 32,
.expected = (unsigned char *)
"\x9c\xdc\x63\x8a\x19\x23\x22\x66\x0c\xc5\xb9\xd7"
"\xfb\x2a\xb0\x31\xe3\x8a\x36\xa8\x5a\xa8\x14\xda"
"\x1e\xa9\xcc\xfe\xb8\x26\x44\x83\x9f\xf6\xff\xaa"
"\xc8\x98\xb8\x30\x35\x3b\x3d\x36\xd2\x49\xd4\x40"
"\x62\x0a\x65\x10\x76\x55\xef\xc0\x95\x9c\xa7\xda"
"\x3f\xcf\xb7\x7b\xc6\xe1\x28\x52\xfc\x0c\xe2\x37"
"\x0d\x83\xa7\x51\x4b\x31\x47\x3c\xe1\x3c\xae\x70"
"\x01\xc8\xa3\xd3\xc2\xac\x77\x9c\xd1\x68\x77\x9b"
"\x58\x27\x3b\xa5\x0f\xc2\x7a\x8b\x04\x65\x62\xd5"
"\xe8\xd6\xfe\x2a\xaf\xd3\xd3\xfe\xbd\x18\xfb\xcd"
"\xcd\x66\xb5\x01\x69\x66\xa0\x3c",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x17\xc1\x56\xcb\xcc\x50\xd6\x03\x7d\x45\x76\xa3"
"\x75\x76\xc1\x4a\x66\x1b\x2e\xdf\xb0\x2e\x7d\x56"
"\x6d\x99\x3b\xc6\x58\xda\x03\xf6",
.addtlb = (unsigned char *)
"\x7c\x7b\x4a\x4b\x32\x5e\x6f\x67\x34\xf5\x21\x4c"
"\xf9\x96\xf9\xbf\x1c\x8c\x81\xd3\x9b\x60\x6a\x44"
"\xc6\x03\xa2\xfb\x13\x20\x19\xb7",
.addtllen = 32,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x13\x54\x96\xfc\x1b\x7d\x28\xf3\x18\xc9\xa7\x89"
"\xb6\xb3\xc8\x72\xac\x00\xd4\x59\x36\x25\x05\xaf"
"\xa5\xdb\x96\xcb\x3c\x58\x46\x87\xa5\xaa\xbf\x20"
"\x3b\xfe\x23\x0e\xd1\xc7\x41\x0f\x3f\xc9\xb3\x67",
.entropylen = 48,
.entpra = (unsigned char *)
"\xe2\xbd\xb7\x48\x08\x06\xf3\xe1\x93\x3c\xac\x79"
"\xa7\x2b\x11\xda\xe3\x2e\xe1\x91\xa5\x02\x19\x57"
"\x20\x28\xad\xf2\x60\xd7\xcd\x45",
.entprb = (unsigned char *)
"\x8b\xd4\x69\xfc\xff\x59\x95\x95\xc6\x51\xde\x71"
"\x68\x5f\xfc\xf9\x4a\xab\xec\x5a\xcb\xbe\xd3\x66"
"\x1f\xfa\x74\xd3\xac\xa6\x74\x60",
.entprlen = 32,
.expected = (unsigned char *)
"\x1f\x9e\xaf\xe4\xd2\x46\xb7\x47\x41\x4c\x65\x99"
"\x01\xe9\x3b\xbb\x83\x0c\x0a\xb0\xc1\x3a\xe2\xb3"
"\x31\x4e\xeb\x93\x73\xee\x0b\x26\xc2\x63\xa5\x75"
"\x45\x99\xd4\x5c\x9f\xa1\xd4\x45\x87\x6b\x20\x61"
"\x40\xea\x78\xa5\x32\xdf\x9e\x66\x17\xaf\xb1\x88"
"\x9e\x2e\x23\xdd\xc1\xda\x13\x97\x88\xa5\xb6\x5e"
"\x90\x14\x4e\xef\x13\xab\x5c\xd9\x2c\x97\x9e\x7c"
"\xd7\xf8\xce\xea\x81\xf5\xcd\x71\x15\x49\x44\xce"
"\x83\xb6\x05\xfb\x7d\x30\xb5\x57\x2c\x31\x4f\xfc"
"\xfe\x80\xb6\xc0\x13\x0c\x5b\x9b\x2e\x8f\x3d\xfc"
"\xc2\xa3\x0c\x11\x1b\x80\x5f\xf3",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\x64\xb6\xfc\x60\xbc\x61\x76\x23\x6d\x3f\x4a\x0f"
"\xe1\xb4\xd5\x20\x9e\x70\xdd\x03\x53\x6d\xbf\xce"
"\xcd\x56\x80\xbc\xb8\x15\xc8\xaa",
.perslen = 32,
}, {
.entropy = (unsigned char *)
"\xc7\xcc\xbc\x67\x7e\x21\x66\x1e\x27\x2b\x63\xdd"
"\x3a\x78\xdc\xdf\x66\x6d\x3f\x24\xae\xcf\x37\x01"
"\xa9\x0d\x89\x8a\xa7\xdc\x81\x58\xae\xb2\x10\x15"
"\x7e\x18\x44\x6d\x13\xea\xdf\x37\x85\xfe\x81\xfb",
.entropylen = 48,
.entpra = (unsigned char *)
"\x7b\xa1\x91\x5b\x3c\x04\xc4\x1b\x1d\x19\x2f\x1a"
"\x18\x81\x60\x3c\x6c\x62\x91\xb7\xe9\xf5\xcb\x96"
"\xbb\x81\x6a\xcc\xb5\xae\x55\xb6",
.entprb = (unsigned char *)
"\x99\x2c\xc7\x78\x7e\x3b\x88\x12\xef\xbe\xd3\xd2"
"\x7d\x2a\xa5\x86\xda\x8d\x58\x73\x4a\x0a\xb2\x2e"
"\xbb\x4c\x7e\xe3\x9a\xb6\x81\xc1",
.entprlen = 32,
.expected = (unsigned char *)
"\x95\x6f\x95\xfc\x3b\xb7\xfe\x3e\xd0\x4e\x1a\x14"
"\x6c\x34\x7f\x7b\x1d\x0d\x63\x5e\x48\x9c\x69\xe6"
"\x46\x07\xd2\x87\xf3\x86\x52\x3d\x98\x27\x5e\xd7"
"\x54\xe7\x75\x50\x4f\xfb\x4d\xfd\xac\x2f\x4b\x77"
"\xcf\x9e\x8e\xcc\x16\xa2\x24\xcd\x53\xde\x3e\xc5"
"\x55\x5d\xd5\x26\x3f\x89\xdf\xca\x8b\x4e\x1e\xb6"
"\x88\x78\x63\x5c\xa2\x63\x98\x4e\x6f\x25\x59\xb1"
"\x5f\x2b\x23\xb0\x4b\xa5\x18\x5d\xc2\x15\x74\x40"
"\x59\x4c\xb4\x1e\xcf\x9a\x36\xfd\x43\xe2\x03\xb8"
"\x59\x91\x30\x89\x2a\xc8\x5a\x43\x23\x7c\x73\x72"
"\xda\x3f\xad\x2b\xba\x00\x6b\xd1",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x18\xe8\x17\xff\xef\x39\xc7\x41\x5c\x73\x03\x03"
"\xf6\x3d\xe8\x5f\xc8\xab\xe4\xab\x0f\xad\xe8\xd6"
"\x86\x88\x55\x28\xc1\x69\xdd\x76",
.addtlb = (unsigned char *)
"\xac\x07\xfc\xbe\x87\x0e\xd3\xea\x1f\x7e\xb8\xe7"
"\x9d\xec\xe8\xe7\xbc\xf3\x18\x25\x77\x35\x4a\xaa"
"\x00\x99\x2a\xdd\x0a\x00\x50\x82",
.addtllen = 32,
.pers = (unsigned char *)
"\xbc\x55\xab\x3c\xf6\x52\xb0\x11\x3d\x7b\x90\xb8"
"\x24\xc9\x26\x4e\x5a\x1e\x77\x0d\x3d\x58\x4a\xda"
"\xd1\x81\xe9\xf8\xeb\x30\x8f\x6f",
.perslen = 32,
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct drbg_testvec drbg_pr_ctr_aes128_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\xd1\x44\xc6\x61\x81\x6d\xca\x9d\x15\x28\x8a\x42"
"\x94\xd7\x28\x9c\x43\x77\x19\x29\x1a\x6d\xc3\xa2",
.entropylen = 24,
.entpra = (unsigned char *)
"\x96\xd8\x9e\x45\x32\xc9\xd2\x08\x7a\x6d\x97\x15"
"\xb4\xec\x80\xb1",
.entprb = (unsigned char *)
"\x8b\xb6\x72\xb5\x24\x0b\x98\x65\x95\x95\xe9\xc9"
"\x28\x07\xeb\xc2",
.entprlen = 16,
.expected = (unsigned char *)
"\x70\x19\xd0\x4c\x45\x78\xd6\x68\xa9\x9a\xaa\xfe"
"\xc1\xdf\x27\x9a\x1c\x0d\x0d\xf7\x24\x75\x46\xcc"
"\x77\x6b\xdf\x89\xc6\x94\xdc\x74\x50\x10\x70\x18"
"\x9b\xdc\x96\xb4\x89\x23\x40\x1a\xce\x09\x87\xce"
"\xd2\xf3\xd5\xe4\x51\x67\x74\x11\x5a\xcc\x8b\x3b"
"\x8a\xf1\x23\xa8",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x8e\x83\xe0\xeb\x37\xea\x3e\x53\x5e\x17\x6e\x77"
"\xbd\xb1\x53\x90\xfc\xdc\xc1\x3c\x9a\x88\x22\x94",
.entropylen = 24,
.entpra = (unsigned char *)
"\x6a\x85\xe7\x37\xc8\xf1\x04\x31\x98\x4f\xc8\x73"
"\x67\xd1\x08\xf8",
.entprb = (unsigned char *)
"\xd7\xa4\x68\xe2\x12\x74\xc3\xd9\xf1\xb7\x05\xbc"
"\xd4\xba\x04\x58",
.entprlen = 16,
.expected = (unsigned char *)
"\x78\xd6\xa6\x70\xff\xd1\x82\xf5\xa2\x88\x7f\x6d"
"\x3d\x8c\x39\xb1\xa8\xcb\x2c\x91\xab\x14\x7e\xbc"
"\x95\x45\x9f\x24\xb8\x20\xac\x21\x23\xdb\x72\xd7"
"\x12\x8d\x48\x95\xf3\x19\x0c\x43\xc6\x19\x45\xfc"
"\x8b\xac\x40\x29\x73\x00\x03\x45\x5e\x12\xff\x0c"
"\xc1\x02\x41\x82",
.expectedlen = 64,
.addtla = (unsigned char *)
"\xa2\xd9\x38\xcf\x8b\x29\x67\x5b\x65\x62\x6f\xe8"
"\xeb\xb3\x01\x76",
.addtlb = (unsigned char *)
"\x59\x63\x1e\x81\x8a\x14\xa8\xbb\xa1\xb8\x41\x25"
"\xd0\x7f\xcc\x43",
.addtllen = 16,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x04\xd9\x49\xa6\xdc\xe8\x6e\xbb\xf1\x08\x77\x2b"
"\x9e\x08\xca\x92\x65\x16\xda\x99\xa2\x59\xf3\xe8",
.entropylen = 24,
.entpra = (unsigned char *)
"\x38\x7e\x3f\x6b\x51\x70\x7b\x20\xec\x53\xd0\x66"
"\xc3\x0f\xe3\xb0",
.entprb = (unsigned char *)
"\xe0\x86\xa6\xaa\x5f\x72\x2f\xad\xf7\xef\x06\xb8"
"\xd6\x9c\x9d\xe8",
.entprlen = 16,
.expected = (unsigned char *)
"\xc9\x0a\xaf\x85\x89\x71\x44\x66\x4f\x25\x0b\x2b"
"\xde\xd8\xfa\xff\x52\x5a\x1b\x32\x5e\x41\x7a\x10"
"\x1f\xef\x1e\x62\x23\xe9\x20\x30\xc9\x0d\xad\x69"
"\xb4\x9c\x5b\xf4\x87\x42\xd5\xae\x5e\x5e\x43\xcc"
"\xd9\xfd\x0b\x93\x4a\xe3\xd4\x06\x37\x36\x0f\x3f"
"\x72\x82\x0c\xcf",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\xbf\xa4\x9a\x8f\x7b\xd8\xb1\x7a\x9d\xfa\x45\xed"
"\x21\x52\xb3\xad",
.perslen = 16,
}, {
.entropy = (unsigned char *)
"\x92\x89\x8f\x31\xfa\x1c\xff\x6d\x18\x2f\x26\x06"
"\x43\xdf\xf8\x18\xc2\xa4\xd9\x72\xc3\xb9\xb6\x97",
.entropylen = 24,
.entpra = (unsigned char *)
"\x20\x72\x8a\x06\xf8\x6f\x8d\xd4\x41\xe2\x72\xb7"
"\xc4\x2c\xe8\x10",
.entprb = (unsigned char *)
"\x3d\xb0\xf0\x94\xf3\x05\x50\x33\x17\x86\x3e\x22"
"\x08\xf7\xa5\x01",
.entprlen = 16,
.expected = (unsigned char *)
"\x5a\x35\x39\x87\x0f\x4d\x22\xa4\x09\x24\xee\x71"
"\xc9\x6f\xac\x72\x0a\xd6\xf0\x88\x82\xd0\x83\x28"
"\x73\xec\x3f\x93\xd8\xab\x45\x23\xf0\x7e\xac\x45"
"\x14\x5e\x93\x9f\xb1\xd6\x76\x43\x3d\xb6\xe8\x08"
"\x88\xf6\xda\x89\x08\x77\x42\xfe\x1a\xf4\x3f\xc4"
"\x23\xc5\x1f\x68",
.expectedlen = 64,
.addtla = (unsigned char *)
"\x1a\x40\xfa\xe3\xcc\x6c\x7c\xa0\xf8\xda\xba\x59"
"\x23\x6d\xad\x1d",
.addtlb = (unsigned char *)
"\x9f\x72\x76\x6c\xc7\x46\xe5\xed\x2e\x53\x20\x12"
"\xbc\x59\x31\x8c",
.addtllen = 16,
.pers = (unsigned char *)
"\xea\x65\xee\x60\x26\x4e\x7e\xb6\x0e\x82\x68\xc4"
"\x37\x3c\x5c\x0b",
.perslen = 16,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* SP800-90A DRBG Test vectors from
* http://csrc.nist.gov/groups/STM/cavp/documents/drbg/drbgtestvectors.zip
*
* Test vectors for DRBG without prediction resistance. All types of DRBGs
* (Hash, HMAC, CTR) are tested with all permutations of use cases (w/ and
* w/o personalization string, w/ and w/o additional input string).
*/
static const struct drbg_testvec drbg_nopr_sha256_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\xa6\x5a\xd0\xf3\x45\xdb\x4e\x0e\xff\xe8\x75\xc3"
"\xa2\xe7\x1f\x42\xc7\x12\x9d\x62\x0f\xf5\xc1\x19"
"\xa9\xef\x55\xf0\x51\x85\xe0\xfb\x85\x81\xf9\x31"
"\x75\x17\x27\x6e\x06\xe9\x60\x7d\xdb\xcb\xcc\x2e",
.entropylen = 48,
.expected = (unsigned char *)
"\xd3\xe1\x60\xc3\x5b\x99\xf3\x40\xb2\x62\x82\x64"
"\xd1\x75\x10\x60\xe0\x04\x5d\xa3\x83\xff\x57\xa5"
"\x7d\x73\xa6\x73\xd2\xb8\xd8\x0d\xaa\xf6\xa6\xc3"
"\x5a\x91\xbb\x45\x79\xd7\x3f\xd0\xc8\xfe\xd1\x11"
"\xb0\x39\x13\x06\x82\x8a\xdf\xed\x52\x8f\x01\x81"
"\x21\xb3\xfe\xbd\xc3\x43\xe7\x97\xb8\x7d\xbb\x63"
"\xdb\x13\x33\xde\xd9\xd1\xec\xe1\x77\xcf\xa6\xb7"
"\x1f\xe8\xab\x1d\xa4\x66\x24\xed\x64\x15\xe5\x1c"
"\xcd\xe2\xc7\xca\x86\xe2\x83\x99\x0e\xea\xeb\x91"
"\x12\x04\x15\x52\x8b\x22\x95\x91\x02\x81\xb0\x2d"
"\xd4\x31\xf4\xc9\xf7\x04\x27\xdf",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x73\xd3\xfb\xa3\x94\x5f\x2b\x5f\xb9\x8f\xf6\x9c"
"\x8a\x93\x17\xae\x19\xc3\x4c\xc3\xd6\xca\xa3\x2d"
"\x16\xfc\x42\xd2\x2d\xd5\x6f\x56\xcc\x1d\x30\xff"
"\x9e\x06\x3e\x09\xce\x58\xe6\x9a\x35\xb3\xa6\x56",
.entropylen = 48,
.expected = (unsigned char *)
"\x71\x7b\x93\x46\x1a\x40\xaa\x35\xa4\xaa\xc5\xe7"
"\x6d\x5b\x5b\x8a\xa0\xdf\x39\x7d\xae\x71\x58\x5b"
"\x3c\x7c\xb4\xf0\x89\xfa\x4a\x8c\xa9\x5c\x54\xc0"
"\x40\xdf\xbc\xce\x26\x81\x34\xf8\xba\x7d\x1c\xe8"
"\xad\x21\xe0\x74\xcf\x48\x84\x30\x1f\xa1\xd5\x4f"
"\x81\x42\x2f\xf4\xdb\x0b\x23\xf8\x73\x27\xb8\x1d"
"\x42\xf8\x44\x58\xd8\x5b\x29\x27\x0a\xf8\x69\x59"
"\xb5\x78\x44\xeb\x9e\xe0\x68\x6f\x42\x9a\xb0\x5b"
"\xe0\x4e\xcb\x6a\xaa\xe2\xd2\xd5\x33\x25\x3e\xe0"
"\x6c\xc7\x6a\x07\xa5\x03\x83\x9f\xe2\x8b\xd1\x1c"
"\x70\xa8\x07\x59\x97\xeb\xf6\xbe",
.expectedlen = 128,
.addtla = (unsigned char *)
"\xf4\xd5\x98\x3d\xa8\xfc\xfa\x37\xb7\x54\x67\x73"
"\xc7\xc3\xdd\x47\x34\x71\x02\x5d\xc1\xa0\xd3\x10"
"\xc1\x8b\xbd\xf5\x66\x34\x6f\xdd",
.addtlb = (unsigned char *)
"\xf7\x9e\x6a\x56\x0e\x73\xe9\xd9\x7a\xd1\x69\xe0"
"\x6f\x8c\x55\x1c\x44\xd1\xce\x6f\x28\xcc\xa4\x4d"
"\xa8\xc0\x85\xd1\x5a\x0c\x59\x40",
.addtllen = 32,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x2a\x85\xa9\x8b\xd0\xda\x83\xd6\xad\xab\x9f\xbb"
"\x54\x31\x15\x95\x1c\x4d\x49\x9f\x6a\x15\xf6\xe4"
"\x15\x50\x88\x06\x29\x0d\xed\x8d\xb9\x6f\x96\xe1"
"\x83\x9f\xf7\x88\xda\x84\xbf\x44\x28\xd9\x1d\xaa",
.entropylen = 48,
.expected = (unsigned char *)
"\x2d\x55\xde\xc9\xed\x05\x47\x07\x3d\x04\xfc\x28"
"\x0f\x92\xf0\x4d\xd8\x00\x32\x47\x0a\x1b\x1c\x4b"
"\xef\xd9\x97\xa1\x17\x67\xda\x26\x6c\xfe\x76\x46"
"\x6f\xbc\x6d\x82\x4e\x83\x8a\x98\x66\x6c\x01\xb6"
"\xe6\x64\xe0\x08\x10\x6f\xd3\x5d\x90\xe7\x0d\x72"
"\xa6\xa7\xe3\xbb\x98\x11\x12\x56\x23\xc2\x6d\xd1"
"\xc8\xa8\x7a\x39\xf3\x34\xe3\xb8\xf8\x66\x00\x77"
"\x7d\xcf\x3c\x3e\xfa\xc9\x0f\xaf\xe0\x24\xfa\xe9"
"\x84\xf9\x6a\x01\xf6\x35\xdb\x5c\xab\x2a\xef\x4e"
"\xac\xab\x55\xb8\x9b\xef\x98\x68\xaf\x51\xd8\x16"
"\xa5\x5e\xae\xf9\x1e\xd2\xdb\xe6",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\xa8\x80\xec\x98\x30\x98\x15\xd2\xc6\xc4\x68\xf1"
"\x3a\x1c\xbf\xce\x6a\x40\x14\xeb\x36\x99\x53\xda"
"\x57\x6b\xce\xa4\x1c\x66\x3d\xbc",
.perslen = 32,
}, {
.entropy = (unsigned char *)
"\x69\xed\x82\xa9\xc5\x7b\xbf\xe5\x1d\x2f\xcb\x7a"
"\xd3\x50\x7d\x96\xb4\xb9\x2b\x50\x77\x51\x27\x74"
"\x33\x74\xba\xf1\x30\xdf\x8e\xdf\x87\x1d\x87\xbc"
"\x96\xb2\xc3\xa7\xed\x60\x5e\x61\x4e\x51\x29\x1a",
.entropylen = 48,
.expected = (unsigned char *)
"\xa5\x71\x24\x31\x11\xfe\x13\xe1\xa8\x24\x12\xfb"
"\x37\xa1\x27\xa5\xab\x77\xa1\x9f\xae\x8f\xaf\x13"
"\x93\xf7\x53\x85\x91\xb6\x1b\xab\xd4\x6b\xea\xb6"
"\xef\xda\x4c\x90\x6e\xef\x5f\xde\xe1\xc7\x10\x36"
"\xd5\x67\xbd\x14\xb6\x89\x21\x0c\xc9\x92\x65\x64"
"\xd0\xf3\x23\xe0\x7f\xd1\xe8\x75\xc2\x85\x06\xea"
"\xca\xc0\xcb\x79\x2d\x29\x82\xfc\xaa\x9a\xc6\x95"
"\x7e\xdc\x88\x65\xba\xec\x0e\x16\x87\xec\xa3\x9e"
"\xd8\x8c\x80\xab\x3a\x64\xe0\xcb\x0e\x45\x98\xdd"
"\x7c\x6c\x6c\x26\x11\x13\xc8\xce\xa9\x47\xa6\x06"
"\x57\xa2\x66\xbb\x2d\x7f\xf3\xc1",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x74\xd3\x6d\xda\xe8\xd6\x86\x5f\x63\x01\xfd\xf2"
"\x7d\x06\x29\x6d\x94\xd1\x66\xf0\xd2\x72\x67\x4e"
"\x77\xc5\x3d\x9e\x03\xe3\xa5\x78",
.addtlb = (unsigned char *)
"\xf6\xb6\x3d\xf0\x7c\x26\x04\xc5\x8b\xcd\x3e\x6a"
"\x9f\x9c\x3a\x2e\xdb\x47\x87\xe5\x8e\x00\x5e\x2b"
"\x74\x7f\xa6\xf6\x80\xcd\x9b\x21",
.addtllen = 32,
.pers = (unsigned char *)
"\x74\xa6\xe0\x08\xf9\x27\xee\x1d\x6e\x3c\x28\x20"
"\x87\xdd\xd7\x54\x31\x47\x78\x4b\xe5\x6d\xa3\x73"
"\xa9\x65\xb1\x10\xc1\xdc\x77\x7c",
.perslen = 32,
},
};
static const struct drbg_testvec drbg_nopr_hmac_sha256_tv_template[] = {
{
.entropy = (unsigned char *)
"\xca\x85\x19\x11\x34\x93\x84\xbf\xfe\x89\xde\x1c"
"\xbd\xc4\x6e\x68\x31\xe4\x4d\x34\xa4\xfb\x93\x5e"
"\xe2\x85\xdd\x14\xb7\x1a\x74\x88\x65\x9b\xa9\x6c"
"\x60\x1d\xc6\x9f\xc9\x02\x94\x08\x05\xec\x0c\xa8",
.entropylen = 48,
.expected = (unsigned char *)
"\xe5\x28\xe9\xab\xf2\xde\xce\x54\xd4\x7c\x7e\x75"
"\xe5\xfe\x30\x21\x49\xf8\x17\xea\x9f\xb4\xbe\xe6"
"\xf4\x19\x96\x97\xd0\x4d\x5b\x89\xd5\x4f\xbb\x97"
"\x8a\x15\xb5\xc4\x43\xc9\xec\x21\x03\x6d\x24\x60"
"\xb6\xf7\x3e\xba\xd0\xdc\x2a\xba\x6e\x62\x4a\xbf"
"\x07\x74\x5b\xc1\x07\x69\x4b\xb7\x54\x7b\xb0\x99"
"\x5f\x70\xde\x25\xd6\xb2\x9e\x2d\x30\x11\xbb\x19"
"\xd2\x76\x76\xc0\x71\x62\xc8\xb5\xcc\xde\x06\x68"
"\x96\x1d\xf8\x68\x03\x48\x2c\xb3\x7e\xd6\xd5\xc0"
"\xbb\x8d\x50\xcf\x1f\x50\xd4\x76\xaa\x04\x58\xbd"
"\xab\xa8\x06\xf4\x8b\xe9\xdc\xb8",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
}, {
.entropy = (unsigned char *)
"\xf9\x7a\x3c\xfd\x91\xfa\xa0\x46\xb9\xe6\x1b\x94"
"\x93\xd4\x36\xc4\x93\x1f\x60\x4b\x22\xf1\x08\x15"
"\x21\xb3\x41\x91\x51\xe8\xff\x06\x11\xf3\xa7\xd4"
"\x35\x95\x35\x7d\x58\x12\x0b\xd1\xe2\xdd\x8a\xed",
.entropylen = 48,
.expected = (unsigned char *)
"\xc6\x87\x1c\xff\x08\x24\xfe\x55\xea\x76\x89\xa5"
"\x22\x29\x88\x67\x30\x45\x0e\x5d\x36\x2d\xa5\xbf"
"\x59\x0d\xcf\x9a\xcd\x67\xfe\xd4\xcb\x32\x10\x7d"
"\xf5\xd0\x39\x69\xa6\x6b\x1f\x64\x94\xfd\xf5\xd6"
"\x3d\x5b\x4d\x0d\x34\xea\x73\x99\xa0\x7d\x01\x16"
"\x12\x6d\x0d\x51\x8c\x7c\x55\xba\x46\xe1\x2f\x62"
"\xef\xc8\xfe\x28\xa5\x1c\x9d\x42\x8e\x6d\x37\x1d"
"\x73\x97\xab\x31\x9f\xc7\x3d\xed\x47\x22\xe5\xb4"
"\xf3\x00\x04\x03\x2a\x61\x28\xdf\x5e\x74\x97\xec"
"\xf8\x2c\xa7\xb0\xa5\x0e\x86\x7e\xf6\x72\x8a\x4f"
"\x50\x9a\x8c\x85\x90\x87\x03\x9c",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x51\x72\x89\xaf\xe4\x44\xa0\xfe\x5e\xd1\xa4\x1d"
"\xbb\xb5\xeb\x17\x15\x00\x79\xbd\xd3\x1e\x29\xcf"
"\x2f\xf3\x00\x34\xd8\x26\x8e\x3b",
.addtlb = (unsigned char *)
"\x88\x02\x8d\x29\xef\x80\xb4\xe6\xf0\xfe\x12\xf9"
"\x1d\x74\x49\xfe\x75\x06\x26\x82\xe8\x9c\x57\x14"
"\x40\xc0\xc9\xb5\x2c\x42\xa6\xe0",
.addtllen = 32,
.pers = NULL,
.perslen = 0,
}, {
.entropy = (unsigned char *)
"\x8d\xf0\x13\xb4\xd1\x03\x52\x30\x73\x91\x7d\xdf"
"\x6a\x86\x97\x93\x05\x9e\x99\x43\xfc\x86\x54\x54"
"\x9e\x7a\xb2\x2f\x7c\x29\xf1\x22\xda\x26\x25\xaf"
"\x2d\xdd\x4a\xbc\xce\x3c\xf4\xfa\x46\x59\xd8\x4e",
.entropylen = 48,
.expected = (unsigned char *)
"\xb9\x1c\xba\x4c\xc8\x4f\xa2\x5d\xf8\x61\x0b\x81"
"\xb6\x41\x40\x27\x68\xa2\x09\x72\x34\x93\x2e\x37"
"\xd5\x90\xb1\x15\x4c\xbd\x23\xf9\x74\x52\xe3\x10"
"\xe2\x91\xc4\x51\x46\x14\x7f\x0d\xa2\xd8\x17\x61"
"\xfe\x90\xfb\xa6\x4f\x94\x41\x9c\x0f\x66\x2b\x28"
"\xc1\xed\x94\xda\x48\x7b\xb7\xe7\x3e\xec\x79\x8f"
"\xbc\xf9\x81\xb7\x91\xd1\xbe\x4f\x17\x7a\x89\x07"
"\xaa\x3c\x40\x16\x43\xa5\xb6\x2b\x87\xb8\x9d\x66"
"\xb3\xa6\x0e\x40\xd4\xa8\xe4\xe9\xd8\x2a\xf6\xd2"
"\x70\x0e\x6f\x53\x5c\xdb\x51\xf7\x5c\x32\x17\x29"
"\x10\x37\x41\x03\x0c\xcc\x3a\x56",
.expectedlen = 128,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\xb5\x71\xe6\x6d\x7c\x33\x8b\xc0\x7b\x76\xad\x37"
"\x57\xbb\x2f\x94\x52\xbf\x7e\x07\x43\x7a\xe8\x58"
"\x1c\xe7\xbc\x7c\x3a\xc6\x51\xa9",
.perslen = 32,
}, {
.entropy = (unsigned char *)
"\xc2\xa5\x66\xa9\xa1\x81\x7b\x15\xc5\xc3\xb7\x78"
"\x17\x7a\xc8\x7c\x24\xe7\x97\xbe\x0a\x84\x5f\x11"
"\xc2\xfe\x39\x9d\xd3\x77\x32\xf2\xcb\x18\x94\xeb"
"\x2b\x97\xb3\xc5\x6e\x62\x83\x29\x51\x6f\x86\xec",
.entropylen = 48,
.expected = (unsigned char *)
"\xb3\xa3\x69\x8d\x77\x76\x99\xa0\xdd\x9f\xa3\xf0"
"\xa9\xfa\x57\x83\x2d\x3c\xef\xac\x5d\xf2\x44\x37"
"\xc6\xd7\x3a\x0f\xe4\x10\x40\xf1\x72\x90\x38\xae"
"\xf1\xe9\x26\x35\x2e\xa5\x9d\xe1\x20\xbf\xb7\xb0"
"\x73\x18\x3a\x34\x10\x6e\xfe\xd6\x27\x8f\xf8\xad"
"\x84\x4b\xa0\x44\x81\x15\xdf\xdd\xf3\x31\x9a\x82"
"\xde\x6b\xb1\x1d\x80\xbd\x87\x1a\x9a\xcd\x35\xc7"
"\x36\x45\xe1\x27\x0f\xb9\xfe\x4f\xa8\x8e\xc0\xe4"
"\x65\x40\x9e\xa0\xcb\xa8\x09\xfe\x2f\x45\xe0\x49"
"\x43\xa2\xe3\x96\xbb\xb7\xdd\x2f\x4e\x07\x95\x30"
"\x35\x24\xcc\x9c\xc5\xea\x54\xa1",
.expectedlen = 128,
.addtla = (unsigned char *)
"\x41\x3d\xd8\x3f\xe5\x68\x35\xab\xd4\x78\xcb\x96"
"\x93\xd6\x76\x35\x90\x1c\x40\x23\x9a\x26\x64\x62"
"\xd3\x13\x3b\x83\xe4\x9c\x82\x0b",
.addtlb = (unsigned char *)
"\xd5\xc4\xa7\x1f\x9d\x6d\x95\xa1\xbe\xdf\x0b\xd2"
"\x24\x7c\x27\x7d\x1f\x84\xa4\xe5\x7a\x4a\x88\x25"
"\xb8\x2a\x2d\x09\x7d\xe6\x3e\xf1",
.addtllen = 32,
.pers = (unsigned char *)
"\x13\xce\x4d\x8d\xd2\xdb\x97\x96\xf9\x41\x56\xc8"
"\xe8\xf0\x76\x9b\x0a\xa1\xc8\x2c\x13\x23\xb6\x15"
"\x36\x60\x3b\xca\x37\xc9\xee\x29",
.perslen = 32,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct drbg_testvec drbg_nopr_ctr_aes192_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\xc3\x5c\x2f\xa2\xa8\x9d\x52\xa1\x1f\xa3\x2a\xa9"
"\x6c\x95\xb8\xf1\xc9\xa8\xf9\xcb\x24\x5a\x8b\x40"
"\xf3\xa6\xe5\xa7\xfb\xd9\xd3\xc6\x8e\x27\x7b\xa9"
"\xac\x9b\xbb\x00",
.entropylen = 40,
.expected = (unsigned char *)
"\x8c\x2e\x72\xab\xfd\x9b\xb8\x28\x4d\xb7\x9e\x17"
"\xa4\x3a\x31\x46\xcd\x76\x94\xe3\x52\x49\xfc\x33"
"\x83\x91\x4a\x71\x17\xf4\x13\x68\xe6\xd4\xf1\x48"
"\xff\x49\xbf\x29\x07\x6b\x50\x15\xc5\x9f\x45\x79"
"\x45\x66\x2e\x3d\x35\x03\x84\x3f\x4a\xa5\xa3\xdf"
"\x9a\x9d\xf1\x0d",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
},
};
static const struct drbg_testvec drbg_nopr_ctr_aes256_tv_template[] = {
{
.entropy = (unsigned char *)
"\x36\x40\x19\x40\xfa\x8b\x1f\xba\x91\xa1\x66\x1f"
"\x21\x1d\x78\xa0\xb9\x38\x9a\x74\xe5\xbc\xcf\xec"
"\xe8\xd7\x66\xaf\x1a\x6d\x3b\x14\x49\x6f\x25\xb0"
"\xf1\x30\x1b\x4f\x50\x1b\xe3\x03\x80\xa1\x37\xeb",
.entropylen = 48,
.expected = (unsigned char *)
"\x58\x62\xeb\x38\xbd\x55\x8d\xd9\x78\xa6\x96\xe6"
"\xdf\x16\x47\x82\xdd\xd8\x87\xe7\xe9\xa6\xc9\xf3"
"\xf1\xfb\xaf\xb7\x89\x41\xb5\x35\xa6\x49\x12\xdf"
"\xd2\x24\xc6\xdc\x74\x54\xe5\x25\x0b\x3d\x97\x16"
"\x5e\x16\x26\x0c\x2f\xaf\x1c\xc7\x73\x5c\xb7\x5f"
"\xb4\xf0\x7e\x1d",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
},
};
static const struct drbg_testvec drbg_nopr_ctr_aes128_tv_template[] = {
{
.entropy = (unsigned char *)
"\x87\xe1\xc5\x32\x99\x7f\x57\xa3\x5c\x28\x6d\xe8"
"\x64\xbf\xf2\x64\xa3\x9e\x98\xdb\x6c\x10\x78\x7f",
.entropylen = 24,
.expected = (unsigned char *)
"\x2c\x14\x7e\x24\x11\x9a\xd8\xd4\xb2\xed\x61\xc1"
"\x53\xd0\x50\xc9\x24\xff\x59\x75\x15\xf1\x17\x3a"
"\x3d\xf4\x4b\x2c\x84\x28\xef\x89\x0e\xb9\xde\xf3"
"\xe4\x78\x04\xb2\xfd\x9b\x35\x7f\xe1\x3f\x8a\x3e"
"\x10\xc8\x67\x0a\xf9\xdf\x2d\x6c\x96\xfb\xb2\xb8"
"\xcb\x2d\xd6\xb0",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = NULL,
.perslen = 0,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.entropy = (unsigned char *)
"\x71\xbd\xce\x35\x42\x7d\x20\xbf\x58\xcf\x17\x74"
"\xce\x72\xd8\x33\x34\x50\x2d\x8f\x5b\x14\xc4\xdd",
.entropylen = 24,
.expected = (unsigned char *)
"\x97\x33\xe8\x20\x12\xe2\x7b\xa1\x46\x8f\xf2\x34"
"\xb3\xc9\xb6\x6b\x20\xb2\x4f\xee\x27\xd8\x0b\x21"
"\x8c\xff\x63\x73\x69\x29\xfb\xf3\x85\xcd\x88\x8e"
"\x43\x2c\x71\x8b\xa2\x55\xd2\x0f\x1d\x7f\xe3\xe1"
"\x2a\xa3\xe9\x2c\x25\x89\xc7\x14\x52\x99\x56\xcc"
"\xc3\xdf\xb3\x81",
.expectedlen = 64,
.addtla = (unsigned char *)
"\x66\xef\x42\xd6\x9a\x8c\x3d\x6d\x4a\x9e\x95\xa6"
"\x91\x4d\x81\x56",
.addtlb = (unsigned char *)
"\xe3\x18\x83\xd9\x4b\x5e\xc4\xcc\xaa\x61\x2f\xbb"
"\x4a\x55\xd1\xc6",
.addtllen = 16,
.pers = NULL,
.perslen = 0,
}, {
.entropy = (unsigned char *)
"\xca\x4b\x1e\xfa\x75\xbd\x69\x36\x38\x73\xb8\xf9"
"\xdb\x4d\x35\x0e\x47\xbf\x6c\x37\x72\xfd\xf7\xa9",
.entropylen = 24,
.expected = (unsigned char *)
"\x59\xc3\x19\x79\x1b\xb1\xf3\x0e\xe9\x34\xae\x6e"
"\x8b\x1f\xad\x1f\x74\xca\x25\x45\x68\xb8\x7f\x75"
"\x12\xf8\xf2\xab\x4c\x23\x01\x03\x05\xe1\x70\xee"
"\x75\xd8\xcb\xeb\x23\x4c\x7a\x23\x6e\x12\x27\xdb"
"\x6f\x7a\xac\x3c\x44\xb7\x87\x4b\x65\x56\x74\x45"
"\x34\x30\x0c\x3d",
.expectedlen = 64,
.addtla = NULL,
.addtlb = NULL,
.addtllen = 0,
.pers = (unsigned char *)
"\xeb\xaa\x60\x2c\x4d\xbe\x33\xff\x1b\xef\xbf\x0a"
"\x0b\xc6\x97\x54",
.perslen = 16,
}, {
.entropy = (unsigned char *)
"\xc0\x70\x1f\x92\x50\x75\x8f\xcd\xf2\xbe\x73\x98"
"\x80\xdb\x66\xeb\x14\x68\xb4\xa5\x87\x9c\x2d\xa6",
.entropylen = 24,
.expected = (unsigned char *)
"\x97\xc0\xc0\xe5\xa0\xcc\xf2\x4f\x33\x63\x48\x8a"
"\xdb\x13\x0a\x35\x89\xbf\x80\x65\x62\xee\x13\x95"
"\x7c\x33\xd3\x7d\xf4\x07\x77\x7a\x2b\x65\x0b\x5f"
"\x45\x5c\x13\xf1\x90\x77\x7f\xc5\x04\x3f\xcc\x1a"
"\x38\xf8\xcd\x1b\xbb\xd5\x57\xd1\x4a\x4c\x2e\x8a"
"\x2b\x49\x1e\x5c",
.expectedlen = 64,
.addtla = (unsigned char *)
"\xf9\x01\xf8\x16\x7a\x1d\xff\xde\x8e\x3c\x83\xe2"
"\x44\x85\xe7\xfe",
.addtlb = (unsigned char *)
"\x17\x1c\x09\x38\xc2\x38\x9f\x97\x87\x60\x55\xb4"
"\x82\x16\x62\x7f",
.addtllen = 16,
.pers = (unsigned char *)
"\x80\x08\xae\xe8\xe9\x69\x40\xc5\x08\x73\xc7\x9f"
"\x8e\xcf\xe0\x02",
.perslen = 16,
},
};
/* Cast5 test vectors from RFC 2144 */
static const struct cipher_testvec cast5_tv_template[] = {
{
.key = "\x01\x23\x45\x67\x12\x34\x56\x78"
"\x23\x45\x67\x89\x34\x56\x78\x9a",
.klen = 16,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ctext = "\x23\x8b\x4f\xe5\x84\x7e\x44\xb2",
.len = 8,
}, {
.key = "\x01\x23\x45\x67\x12\x34\x56\x78"
"\x23\x45",
.klen = 10,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ctext = "\xeb\x6a\x71\x1a\x2c\x02\x27\x1b",
.len = 8,
}, {
.key = "\x01\x23\x45\x67\x12",
.klen = 5,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ctext = "\x7a\xc8\x16\xd1\x6e\x9b\x30\x2e",
.len = 8,
}, { /* Generated from TF test vectors */
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A",
.klen = 16,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\x8D\xFC\x81\x9C\xCB\xAA\x5A\x1C"
"\x7E\x95\xCF\x40\xAB\x4D\x6F\xEA"
"\xD3\xD9\xB0\x9A\xB7\xC7\xE0\x2E"
"\xD1\x39\x34\x92\x8F\xFA\x14\xF1"
"\xD5\xD2\x7B\x59\x1F\x35\x28\xC2"
"\x20\xD9\x42\x06\xC9\x0B\x10\x04"
"\xF8\x79\xCD\x32\x86\x75\x4C\xB6"
"\x7B\x1C\x52\xB1\x91\x64\x22\x4B"
"\x13\xC7\xAE\x98\x0E\xB5\xCF\x6F"
"\x3F\xF4\x43\x96\x73\x0D\xA2\x05"
"\xDB\xFD\x28\x90\x2C\x56\xB9\x37"
"\x5B\x69\x0C\xAD\x84\x67\xFF\x15"
"\x4A\xD4\xA7\xD3\xDD\x99\x47\x3A"
"\xED\x34\x35\x78\x6B\x91\xC9\x32"
"\xE1\xBF\xBC\xB4\x04\x85\x6A\x39"
"\xC0\xBA\x51\xD0\x0F\x4E\xD1\xE2"
"\x1C\xFD\x0E\x05\x07\xF4\x10\xED"
"\xA2\x17\xFF\xF5\x64\xC6\x1A\x22"
"\xAD\x78\xE7\xD7\x11\xE9\x99\xB9"
"\xAA\xEC\x6F\xF8\x3B\xBF\xCE\x77"
"\x93\xE8\xAD\x1D\x50\x6C\xAE\xBC"
"\xBA\x5C\x80\xD1\x91\x65\x51\x1B"
"\xE8\x0A\xCD\x99\x96\x71\x3D\xB6"
"\x78\x75\x37\x55\xC1\xF5\x90\x40"
"\x34\xF4\x7E\xC8\xCC\x3A\x5F\x6E"
"\x36\xA1\xA1\xC2\x3A\x72\x42\x8E"
"\x0E\x37\x88\xE8\xCE\x83\xCB\xAD"
"\xE0\x69\x77\x50\xC7\x0C\x99\xCA"
"\x19\x5B\x30\x25\x9A\xEF\x9B\x0C"
"\xEF\x8F\x74\x4C\xCF\x49\x4E\xB9"
"\xC5\xAE\x9E\x2E\x78\x9A\xB9\x48"
"\xD5\x81\xE4\x37\x1D\xBF\x27\xD9"
"\xC5\xD6\x65\x43\x45\x8C\xBB\xB6"
"\x55\xF4\x06\xBB\x49\x53\x8B\x1B"
"\x07\xA9\x96\x69\x5B\xCB\x0F\xBC"
"\x93\x85\x90\x0F\x0A\x68\x40\x2A"
"\x95\xED\x2D\x88\xBF\x71\xD0\xBB"
"\xEC\xB0\x77\x6C\x79\xFC\x3C\x05"
"\x49\x3F\xB8\x24\xEF\x8E\x09\xA2"
"\x1D\xEF\x92\x02\x96\xD4\x7F\xC8"
"\x03\xB2\xCA\xDB\x17\x5C\x52\xCF"
"\xDD\x70\x37\x63\xAA\xA5\x83\x20"
"\x52\x02\xF6\xB9\xE7\x6E\x0A\xB6"
"\x79\x03\xA0\xDA\xA3\x79\x21\xBD"
"\xE3\x37\x3A\xC0\xF7\x2C\x32\xBE"
"\x8B\xE8\xA6\x00\xC7\x32\xD5\x06"
"\xBB\xE3\xAB\x06\x21\x82\xB8\x32"
"\x31\x34\x2A\xA7\x1F\x64\x99\xBF"
"\xFA\xDA\x3D\x75\xF7\x48\xD5\x48"
"\x4B\x52\x7E\xF6\x7C\xAB\x67\x59"
"\xC5\xDC\xA8\xC6\x63\x85\x4A\xDF"
"\xF0\x40\x5F\xCF\xE3\x58\x52\x67"
"\x7A\x24\x32\xC5\xEC\x9E\xA9\x6F"
"\x58\x56\xDD\x94\x1F\x71\x8D\xF4"
"\x6E\xFF\x2C\xA7\xA5\xD8\xBA\xAF"
"\x1D\x8B\xA2\x46\xB5\xC4\x9F\x57"
"\x8D\xD8\xB3\x3C\x02\x0D\xBB\x84"
"\xC7\xBD\xB4\x9A\x6E\xBB\xB1\x37"
"\x95\x79\xC4\xA7\xEA\x1D\xDC\x33"
"\x5D\x0B\x3F\x03\x8F\x30\xF9\xAE"
"\x4F\xFE\x24\x9C\x9A\x02\xE5\x57"
"\xF5\xBC\x25\xD6\x02\x56\x57\x1C",
.len = 496,
.also_non_np = 1,
.np = 3,
.tap = { 496 - 20, 4, 16 },
},
};
static const struct cipher_testvec cast5_cbc_tv_template[] = {
{ /* Generated from TF test vectors */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A",
.klen = 16,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\x05\x28\xCE\x61\x90\x80\xE1\x78"
"\xB9\x2A\x97\x7C\xB0\x83\xD8\x1A"
"\xDE\x58\x7F\xD7\xFD\x72\xB8\xFB"
"\xDA\xF0\x6E\x77\x14\x47\x82\xBA"
"\x29\x0E\x25\x6E\xB4\x39\xD9\x7F"
"\x05\xA7\xA7\x3A\xC1\x5D\x9E\x39"
"\xA7\xFB\x0D\x05\x00\xF3\x58\x67"
"\x60\xEC\x73\x77\x46\x85\x9B\x6A"
"\x08\x3E\xBE\x59\xFB\xE4\x96\x34"
"\xB4\x05\x49\x1A\x97\x43\xAD\xA0"
"\xA9\x1E\x6E\x74\xF1\x94\xEC\xA8"
"\xB5\x8A\x20\xEA\x89\x6B\x19\xAA"
"\xA7\xF1\x33\x67\x90\x23\x0D\xEE"
"\x81\xD5\x78\x4F\xD3\x63\xEA\x46"
"\xB5\xB2\x6E\xBB\xCA\x76\x06\x10"
"\x96\x2A\x0A\xBA\xF9\x41\x5A\x1D"
"\x36\x7C\x56\x14\x54\x83\xFA\xA1"
"\x27\xDD\xBA\x8A\x90\x29\xD6\xA6"
"\xFA\x48\x3E\x1E\x23\x6E\x98\xA8"
"\xA7\xD9\x67\x92\x5C\x13\xB4\x71"
"\xA8\xAA\x89\x4A\xA4\xB3\x49\x7C"
"\x7D\x7F\xCE\x6F\x29\x2E\x7E\x37"
"\xC8\x52\x60\xD9\xE7\xCA\x60\x98"
"\xED\xCD\xE8\x60\x83\xAD\x34\x4D"
"\x96\x4A\x99\x2B\xB7\x14\x75\x66"
"\x6C\x2C\x1A\xBA\x4B\xBB\x49\x56"
"\xE1\x86\xA2\x0E\xD0\xF0\x07\xD3"
"\x18\x38\x09\x9C\x0E\x8B\x86\x07"
"\x90\x12\x37\x49\x27\x98\x69\x18"
"\xB0\xCC\xFB\xD3\xBD\x04\xA0\x85"
"\x4B\x22\x97\x07\xB6\x97\xE9\x95"
"\x0F\x88\x36\xA9\x44\x00\xC6\xE9"
"\x27\x53\x5C\x5B\x1F\xD3\xE2\xEE"
"\xD0\xCD\x63\x30\xA9\xC0\xDD\x49"
"\xFE\x16\xA4\x07\x0D\xE2\x5D\x97"
"\xDE\x89\xBA\x2E\xF3\xA9\x5E\xBE"
"\x03\x55\x0E\x02\x41\x4A\x45\x06"
"\xBE\xEA\x32\xF2\xDC\x91\x5C\x20"
"\x94\x02\x30\xD2\xFC\x29\xFA\x8E"
"\x34\xA0\x31\xB8\x34\xBA\xAE\x54"
"\xB5\x88\x1F\xDC\x43\xDC\x22\x9F"
"\xDC\xCE\xD3\xFA\xA4\xA8\xBC\x8A"
"\xC7\x5A\x43\x21\xA5\xB1\xDB\xC3"
"\x84\x3B\xB4\x9B\xB5\xA7\xF1\x0A"
"\xB6\x37\x21\x19\x55\xC2\xBD\x99"
"\x49\x24\xBB\x7C\xB3\x8E\xEF\xD2"
"\x3A\xCF\xA0\x31\x28\x0E\x25\xA2"
"\x11\xB4\x18\x17\x1A\x65\x92\x56"
"\xE8\xE0\x52\x9C\x61\x18\x2A\xB1"
"\x1A\x01\x22\x45\x17\x62\x52\x6C"
"\x91\x44\xCF\x98\xC7\xC0\x79\x26"
"\x32\x66\x6F\x23\x7F\x94\x36\x88"
"\x3C\xC9\xD0\xB7\x45\x30\x31\x86"
"\x3D\xC6\xA3\x98\x62\x84\x1A\x8B"
"\x16\x88\xC7\xA3\xE9\x4F\xE0\x86"
"\xA4\x93\xA8\x34\x5A\xCA\xDF\xCA"
"\x46\x38\xD2\xF4\xE0\x2D\x1E\xC9"
"\x7C\xEF\x53\xB7\x60\x72\x41\xBF"
"\x29\x00\x87\x02\xAF\x44\x4C\xB7"
"\x8C\xF5\x3F\x19\xF4\x80\x45\xA7"
"\x15\x5F\xDB\xE9\xB1\x83\xD2\xE6"
"\x1D\x18\x66\x44\x5B\x8F\x14\xEB",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.tap = { 496 - 20, 4, 16 },
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cast5_ctr_tv_template[] = {
{ /* Generated from TF test vectors */
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A",
.klen = 16,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A",
.ctext = "\xFF\xC4\x2E\x82\x3D\xF8\xA8\x39"
"\x7C\x52\xC4\xD3\xBB\x62\xC6\xA8"
"\x0C",
.len = 17,
}, { /* Generated from TF test vectors */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A",
.klen = 16,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\xFF\xC4\x2E\x82\x3D\xF8\xA8\x39"
"\x7C\x52\xC4\xD3\xBB\x62\xC6\xA8"
"\x0C\x63\xA5\x55\xE3\xF8\x1C\x7F"
"\xDC\x59\xF9\xA0\x52\xAD\x83\xDF"
"\xD5\x3B\x53\x4A\xAA\x1F\x49\x44"
"\xE8\x20\xCC\xF8\x97\xE6\xE0\x3C"
"\x5A\xD2\x83\xEC\xEE\x25\x3F\xCF"
"\x0D\xC2\x79\x80\x99\x6E\xFF\x7B"
"\x64\xB0\x7B\x86\x29\x1D\x9F\x17"
"\x10\xA5\xA5\xEB\x16\x55\x9E\xE3"
"\x88\x18\x52\x56\x48\x58\xD1\x6B"
"\xE8\x74\x6E\x48\xB0\x2E\x69\x63"
"\x32\xAA\xAC\x26\x55\x45\x94\xDE"
"\x30\x26\x26\xE6\x08\x82\x2F\x5F"
"\xA7\x15\x94\x07\x75\x2D\xC6\x3A"
"\x1B\xA0\x39\xFB\xBA\xB9\x06\x56"
"\xF6\x9F\xF1\x2F\x9B\xF3\x89\x8B"
"\x08\xC8\x9D\x5E\x6B\x95\x09\xC7"
"\x98\xB7\x62\xA4\x1D\x25\xFA\xC5"
"\x62\xC8\x5D\x6B\xB4\x85\x88\x7F"
"\x3B\x29\xF9\xB4\x32\x62\x69\xBF"
"\x32\xB8\xEB\xFD\x0E\x26\xAA\xA3"
"\x44\x67\x90\x20\xAC\x41\xDF\x43"
"\xC6\xC7\x19\x9F\x2C\x28\x74\xEB"
"\x3E\x7F\x7A\x80\x5B\xE4\x08\x60"
"\xC7\xC9\x71\x34\x44\xCE\x05\xFD"
"\xA8\x91\xA8\x44\x5E\xD3\x89\x2C"
"\xAE\x59\x0F\x07\x88\x79\x53\x26"
"\xAF\xAC\xCB\x1D\x6F\x08\x25\x62"
"\xD0\x82\x65\x66\xE4\x2A\x29\x1C"
"\x9C\x64\x5F\x49\x9D\xF8\x62\xF9"
"\xED\xC4\x13\x52\x75\xDC\xE4\xF9"
"\x68\x0F\x8A\xCD\xA6\x8D\x75\xAA"
"\x49\xA1\x86\x86\x37\x5C\x6B\x3D"
"\x56\xE5\x6F\xBE\x27\xC0\x10\xF8"
"\x3C\x4D\x17\x35\x14\xDC\x1C\xA0"
"\x6E\xAE\xD1\x10\xDD\x83\x06\xC2"
"\x23\xD3\xC7\x27\x15\x04\x2C\x27"
"\xDD\x1F\x2E\x97\x09\x9C\x33\x7D"
"\xAC\x50\x1B\x2E\xC9\x52\x0C\x14"
"\x4B\x78\xC4\xDE\x07\x6A\x12\x02"
"\x6E\xD7\x4B\x91\xB9\x88\x4D\x02"
"\xC3\xB5\x04\xBC\xE0\x67\xCA\x18"
"\x22\xA1\xAE\x9A\x21\xEF\xB2\x06"
"\x35\xCD\xEC\x37\x70\x2D\xFC\x1E"
"\xA8\x31\xE7\xFC\xE5\x8E\x88\x66"
"\x16\xB5\xC8\x45\x21\x37\xBD\x24"
"\xA9\xD5\x36\x12\x9F\x6E\x67\x80"
"\x87\x54\xD5\xAF\x97\xE1\x15\xA7"
"\x11\xF0\x63\x7B\xE1\x44\x14\x1C"
"\x06\x32\x05\x8C\x6C\xDB\x9B\x36"
"\x6A\x6B\xAD\x3A\x27\x55\x20\x4C"
"\x76\x36\x43\xE8\x16\x60\xB5\xF3"
"\xDF\x5A\xC6\xA5\x69\x78\x59\x51"
"\x54\x68\x65\x06\x84\xDE\x3D\xAE"
"\x38\x91\xBD\xCC\xA2\x8A\xEC\xE6"
"\x9E\x83\xAE\x1E\x8E\x34\x5D\xDE"
"\x91\xCE\x8F\xED\x40\xF7\xC8\x8B"
"\x9A\x13\x4C\xAD\x89\x97\x9E\xD1"
"\x91\x01\xD7\x21\x23\x28\x1E\xCC"
"\x8C\x98\xDB\xDE\xFC\x72\x94\xAA"
"\xC0\x0D\x96\xAA\x23\xF8\xFE\x13",
.len = 496,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.tap = { 496 - 20, 4, 16 },
},
};
/*
* ARC4 test vectors from OpenSSL
*/
static const struct cipher_testvec arc4_tv_template[] = {
{
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.ctext = "\x75\xb7\x87\x80\x99\xe0\xc5\x96",
.len = 8,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x74\x94\xc2\xe7\x10\x4b\x08\x79",
.len = 8,
}, {
.key = "\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 8,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\xde\x18\x89\x41\xa3\x37\x5d\x3a",
.len = 8,
}, {
.key = "\xef\x01\x23\x45",
.klen = 4,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00",
.ctext = "\xd6\xa1\x41\xa7\xec\x3c\x38\xdf"
"\xbd\x61\x5a\x11\x62\xe1\xc7\xba"
"\x36\xb6\x78\x58",
.len = 20,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef",
.klen = 8,
.ptext = "\x12\x34\x56\x78\x9A\xBC\xDE\xF0"
"\x12\x34\x56\x78\x9A\xBC\xDE\xF0"
"\x12\x34\x56\x78\x9A\xBC\xDE\xF0"
"\x12\x34\x56\x78",
.ctext = "\x66\xa0\x94\x9f\x8a\xf7\xd6\x89"
"\x1f\x7f\x83\x2b\xa8\x33\xc0\x0c"
"\x89\x2e\xbe\x30\x14\x3c\xe2\x87"
"\x40\x01\x1e\xcf",
.len = 28,
}, {
.key = "\xef\x01\x23\x45",
.klen = 4,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00",
.ctext = "\xd6\xa1\x41\xa7\xec\x3c\x38\xdf"
"\xbd\x61",
.len = 10,
}, {
.key = "\x01\x23\x45\x67\x89\xAB\xCD\xEF"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.ptext = "\x01\x23\x45\x67\x89\xAB\xCD\xEF",
.ctext = "\x69\x72\x36\x59\x1B\x52\x42\xB1",
.len = 8,
},
};
/*
* TEA test vectors
*/
static const struct cipher_testvec tea_tv_template[] = {
{
.key = zeroed_string,
.klen = 16,
.ptext = zeroed_string,
.ctext = "\x0a\x3a\xea\x41\x40\xa9\xba\x94",
.len = 8,
}, {
.key = "\x2b\x02\x05\x68\x06\x14\x49\x76"
"\x77\x5d\x0e\x26\x6c\x28\x78\x43",
.klen = 16,
.ptext = "\x74\x65\x73\x74\x20\x6d\x65\x2e",
.ctext = "\x77\x5d\x2a\x6a\xf6\xce\x92\x09",
.len = 8,
}, {
.key = "\x09\x65\x43\x11\x66\x44\x39\x25"
"\x51\x3a\x16\x10\x0a\x08\x12\x6e",
.klen = 16,
.ptext = "\x6c\x6f\x6e\x67\x65\x72\x5f\x74"
"\x65\x73\x74\x5f\x76\x65\x63\x74",
.ctext = "\xbe\x7a\xbb\x81\x95\x2d\x1f\x1e"
"\xdd\x89\xa1\x25\x04\x21\xdf\x95",
.len = 16,
}, {
.key = "\x4d\x76\x32\x17\x05\x3f\x75\x2c"
"\x5d\x04\x16\x36\x15\x72\x63\x2f",
.klen = 16,
.ptext = "\x54\x65\x61\x20\x69\x73\x20\x67"
"\x6f\x6f\x64\x20\x66\x6f\x72\x20"
"\x79\x6f\x75\x21\x21\x21\x20\x72"
"\x65\x61\x6c\x6c\x79\x21\x21\x21",
.ctext = "\xe0\x4d\x5d\x3c\xb7\x8c\x36\x47"
"\x94\x18\x95\x91\xa9\xfc\x49\xf8"
"\x44\xd1\x2d\xc2\x99\xb8\x08\x2a"
"\x07\x89\x73\xc2\x45\x92\xc6\x90",
.len = 32,
}
};
/*
* XTEA test vectors
*/
static const struct cipher_testvec xtea_tv_template[] = {
{
.key = zeroed_string,
.klen = 16,
.ptext = zeroed_string,
.ctext = "\xd8\xd4\xe9\xde\xd9\x1e\x13\xf7",
.len = 8,
}, {
.key = "\x2b\x02\x05\x68\x06\x14\x49\x76"
"\x77\x5d\x0e\x26\x6c\x28\x78\x43",
.klen = 16,
.ptext = "\x74\x65\x73\x74\x20\x6d\x65\x2e",
.ctext = "\x94\xeb\xc8\x96\x84\x6a\x49\xa8",
.len = 8,
}, {
.key = "\x09\x65\x43\x11\x66\x44\x39\x25"
"\x51\x3a\x16\x10\x0a\x08\x12\x6e",
.klen = 16,
.ptext = "\x6c\x6f\x6e\x67\x65\x72\x5f\x74"
"\x65\x73\x74\x5f\x76\x65\x63\x74",
.ctext = "\x3e\xce\xae\x22\x60\x56\xa8\x9d"
"\x77\x4d\xd4\xb4\x87\x24\xe3\x9a",
.len = 16,
}, {
.key = "\x4d\x76\x32\x17\x05\x3f\x75\x2c"
"\x5d\x04\x16\x36\x15\x72\x63\x2f",
.klen = 16,
.ptext = "\x54\x65\x61\x20\x69\x73\x20\x67"
"\x6f\x6f\x64\x20\x66\x6f\x72\x20"
"\x79\x6f\x75\x21\x21\x21\x20\x72"
"\x65\x61\x6c\x6c\x79\x21\x21\x21",
.ctext = "\x99\x81\x9f\x5d\x6f\x4b\x31\x3a"
"\x86\xff\x6f\xd0\xe3\x87\x70\x07"
"\x4d\xb8\xcf\xf3\x99\x50\xb3\xd4"
"\x73\xa2\xfa\xc9\x16\x59\x5d\x81",
.len = 32,
}
};
/*
* KHAZAD test vectors.
*/
static const struct cipher_testvec khazad_tv_template[] = {
{
.key = "\x80\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x49\xa4\xce\x32\xac\x19\x0e\x3f",
.len = 8,
}, {
.key = "\x38\x38\x38\x38\x38\x38\x38\x38"
"\x38\x38\x38\x38\x38\x38\x38\x38",
.klen = 16,
.ptext = "\x38\x38\x38\x38\x38\x38\x38\x38",
.ctext = "\x7e\x82\x12\xa1\xd9\x5b\xe4\xf9",
.len = 8,
}, {
.key = "\xa2\xa2\xa2\xa2\xa2\xa2\xa2\xa2"
"\xa2\xa2\xa2\xa2\xa2\xa2\xa2\xa2",
.klen = 16,
.ptext = "\xa2\xa2\xa2\xa2\xa2\xa2\xa2\xa2",
.ctext = "\xaa\xbe\xc1\x95\xc5\x94\x1a\x9c",
.len = 8,
}, {
.key = "\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f"
"\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f",
.klen = 16,
.ptext = "\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f",
.ctext = "\x04\x74\xf5\x70\x50\x16\xd3\xb8",
.len = 8,
}, {
.key = "\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f"
"\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f",
.klen = 16,
.ptext = "\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f"
"\x2f\x2f\x2f\x2f\x2f\x2f\x2f\x2f",
.ctext = "\x04\x74\xf5\x70\x50\x16\xd3\xb8"
"\x04\x74\xf5\x70\x50\x16\xd3\xb8",
.len = 16,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
/*
* Anubis test vectors.
*/
static const struct cipher_testvec anubis_tv_template[] = {
{
.key = "\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe",
.klen = 16,
.ptext = "\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe",
.ctext = "\x6d\xc5\xda\xa2\x26\x7d\x62\x6f"
"\x08\xb7\x52\x8e\x6e\x6e\x86\x90",
.len = 16,
}, {
.key = "\x03\x03\x03\x03\x03\x03\x03\x03"
"\x03\x03\x03\x03\x03\x03\x03\x03"
"\x03\x03\x03\x03",
.klen = 20,
.ptext = "\x03\x03\x03\x03\x03\x03\x03\x03"
"\x03\x03\x03\x03\x03\x03\x03\x03",
.ctext = "\xdb\xf1\x42\xf4\xd1\x8a\xc7\x49"
"\x87\x41\x6f\x82\x0a\x98\x64\xae",
.len = 16,
}, {
.key = "\x24\x24\x24\x24\x24\x24\x24\x24"
"\x24\x24\x24\x24\x24\x24\x24\x24"
"\x24\x24\x24\x24\x24\x24\x24\x24"
"\x24\x24\x24\x24",
.klen = 28,
.ptext = "\x24\x24\x24\x24\x24\x24\x24\x24"
"\x24\x24\x24\x24\x24\x24\x24\x24",
.ctext = "\xfd\x1b\x4a\xe3\xbf\xf0\xad\x3d"
"\x06\xd3\x61\x27\xfd\x13\x9e\xde",
.len = 16,
}, {
.key = "\x25\x25\x25\x25\x25\x25\x25\x25"
"\x25\x25\x25\x25\x25\x25\x25\x25"
"\x25\x25\x25\x25\x25\x25\x25\x25"
"\x25\x25\x25\x25\x25\x25\x25\x25",
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x25\x25\x25\x25\x25\x25\x25\x25"
"\x25\x25\x25\x25\x25\x25\x25\x25",
.ctext = "\x1a\x91\xfb\x2b\xb7\x78\x6b\xc4"
"\x17\xd9\xff\x40\x3b\x0e\xe5\xfe",
.len = 16,
}, {
.key = "\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35",
.klen = 40,
.ptext = "\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35",
.ctext = "\xa5\x2c\x85\x6f\x9c\xba\xa0\x97"
"\x9e\xc6\x84\x0f\x17\x21\x07\xee",
.len = 16,
},
};
static const struct cipher_testvec anubis_cbc_tv_template[] = {
{
.key = "\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe",
.klen = 16,
.ptext = "\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe"
"\xfe\xfe\xfe\xfe\xfe\xfe\xfe\xfe",
.ctext = "\x6d\xc5\xda\xa2\x26\x7d\x62\x6f"
"\x08\xb7\x52\x8e\x6e\x6e\x86\x90"
"\x86\xd8\xb5\x6f\x98\x5e\x8a\x66"
"\x4f\x1f\x78\xa1\xbb\x37\xf1\xbe",
.len = 32,
}, {
.key = "\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35",
.klen = 40,
.ptext = "\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35"
"\x35\x35\x35\x35\x35\x35\x35\x35",
.ctext = "\xa5\x2c\x85\x6f\x9c\xba\xa0\x97"
"\x9e\xc6\x84\x0f\x17\x21\x07\xee"
"\xa2\xbc\x06\x98\xc6\x4b\xda\x75"
"\x2e\xaa\xbe\x58\xce\x01\x5b\xc7",
.len = 32,
},
};
/*
* XETA test vectors
*/
static const struct cipher_testvec xeta_tv_template[] = {
{
.key = zeroed_string,
.klen = 16,
.ptext = zeroed_string,
.ctext = "\xaa\x22\x96\xe5\x6c\x61\xf3\x45",
.len = 8,
}, {
.key = "\x2b\x02\x05\x68\x06\x14\x49\x76"
"\x77\x5d\x0e\x26\x6c\x28\x78\x43",
.klen = 16,
.ptext = "\x74\x65\x73\x74\x20\x6d\x65\x2e",
.ctext = "\x82\x3e\xeb\x35\xdc\xdd\xd9\xc3",
.len = 8,
}, {
.key = "\x09\x65\x43\x11\x66\x44\x39\x25"
"\x51\x3a\x16\x10\x0a\x08\x12\x6e",
.klen = 16,
.ptext = "\x6c\x6f\x6e\x67\x65\x72\x5f\x74"
"\x65\x73\x74\x5f\x76\x65\x63\x74",
.ctext = "\xe2\x04\xdb\xf2\x89\x85\x9e\xea"
"\x61\x35\xaa\xed\xb5\xcb\x71\x2c",
.len = 16,
}, {
.key = "\x4d\x76\x32\x17\x05\x3f\x75\x2c"
"\x5d\x04\x16\x36\x15\x72\x63\x2f",
.klen = 16,
.ptext = "\x54\x65\x61\x20\x69\x73\x20\x67"
"\x6f\x6f\x64\x20\x66\x6f\x72\x20"
"\x79\x6f\x75\x21\x21\x21\x20\x72"
"\x65\x61\x6c\x6c\x79\x21\x21\x21",
.ctext = "\x0b\x03\xcd\x8a\xbe\x95\xfd\xb1"
"\xc1\x44\x91\x0b\xa5\xc9\x1b\xb4"
"\xa9\xda\x1e\x9e\xb1\x3e\x2a\x8f"
"\xea\xa5\x6a\x85\xd1\xf4\xa8\xa5",
.len = 32,
}
};
/*
* FCrypt test vectors
*/
static const struct cipher_testvec fcrypt_pcbc_tv_template[] = {
{ /* http://www.openafs.org/pipermail/openafs-devel/2000-December/005320.html */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 8,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ctext = "\x0E\x09\x00\xC7\x3E\xF7\xED\x41",
.len = 8,
}, {
.key = "\x11\x44\x77\xAA\xDD\x00\x33\x66",
.klen = 8,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ptext = "\x12\x34\x56\x78\x9A\xBC\xDE\xF0",
.ctext = "\xD8\xED\x78\x74\x77\xEC\x06\x80",
.len = 8,
}, { /* From Arla */
.key = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87",
.klen = 8,
.iv = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ptext = "The quick brown fox jumps over the lazy dogs.\0\0",
.ctext = "\x00\xf0\x0e\x11\x75\xe6\x23\x82"
"\xee\xac\x98\x62\x44\x51\xe4\x84"
"\xc3\x59\xd8\xaa\x64\x60\xae\xf7"
"\xd2\xd9\x13\x79\x72\xa3\x45\x03"
"\x23\xb5\x62\xd7\x0c\xf5\x27\xd1"
"\xf8\x91\x3c\xac\x44\x22\x92\xef",
.len = 48,
}, {
.key = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.klen = 8,
.iv = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87",
.ptext = "The quick brown fox jumps over the lazy dogs.\0\0",
.ctext = "\xca\x90\xf5\x9d\xcb\xd4\xd2\x3c"
"\x01\x88\x7f\x3e\x31\x6e\x62\x9d"
"\xd8\xe0\x57\xa3\x06\x3a\x42\x58"
"\x2a\x28\xfe\x72\x52\x2f\xdd\xe0"
"\x19\x89\x09\x1c\x2a\x8e\x8c\x94"
"\xfc\xc7\x68\xe4\x88\xaa\xde\x0f",
.len = 48,
}, { /* split-page version */
.key = "\xfe\xdc\xba\x98\x76\x54\x32\x10",
.klen = 8,
.iv = "\xf0\xe1\xd2\xc3\xb4\xa5\x96\x87",
.ptext = "The quick brown fox jumps over the lazy dogs.\0\0",
.ctext = "\xca\x90\xf5\x9d\xcb\xd4\xd2\x3c"
"\x01\x88\x7f\x3e\x31\x6e\x62\x9d"
"\xd8\xe0\x57\xa3\x06\x3a\x42\x58"
"\x2a\x28\xfe\x72\x52\x2f\xdd\xe0"
"\x19\x89\x09\x1c\x2a\x8e\x8c\x94"
"\xfc\xc7\x68\xe4\x88\xaa\xde\x0f",
.len = 48,
.np = 2,
.tap = { 20, 28 },
}
};
/*
* CAMELLIA test vectors.
*/
static const struct cipher_testvec camellia_tv_template[] = {
{
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10",
.klen = 16,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\x67\x67\x31\x38\x54\x96\x69\x73"
"\x08\x57\x06\x56\x48\xea\xbe\x43",
.len = 16,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10"
"\x00\x11\x22\x33\x44\x55\x66\x77",
.klen = 24,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\xb4\x99\x34\x01\xb3\xe9\x96\xf8"
"\x4e\xe5\xce\xe7\xd7\x9b\x09\xb9",
.len = 16,
}, {
.key = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10"
"\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff",
.klen = 32,
.ptext = "\x01\x23\x45\x67\x89\xab\xcd\xef"
"\xfe\xdc\xba\x98\x76\x54\x32\x10",
.ctext = "\x9a\xcc\x23\x7d\xff\x16\xd7\x6c"
"\x20\xef\x7c\x91\x9e\x3a\x75\x09",
.len = 16,
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
}, { /* Generated with Crypto++ */
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.key = "\x3F\x85\x62\x3F\x1C\xF9\xD6\x1C"
"\xF9\xD6\xB3\x90\x6D\x4A\x90\x6D"
"\x4A\x27\x04\xE1\x27\x04\xE1\xBE"
"\x9B\x78\xBE\x9B\x78\x55\x32\x0F",
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2 Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-13 04:46:35 -06:00
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06"
"\x9D\x34\xCB\x3F\xD6\x6D\x04\x78"
"\x0F\xA6\x1A\xB1\x48\xDF\x53\xEA"
"\x81\x18\x8C\x23\xBA\x2E\xC5\x5C"
"\xF3\x67\xFE\x95\x09\xA0\x37\xCE"
"\x42\xD9\x70\x07\x7B\x12\xA9\x1D"
"\xB4\x4B\xE2\x56\xED\x84\x1B\x8F"
"\x26\xBD\x31\xC8\x5F\xF6\x6A\x01"
"\x98\x0C\xA3\x3A\xD1\x45\xDC\x73"
"\x0A\x7E\x15\xAC\x20\xB7\x4E\xE5"
"\x59\xF0\x87\x1E\x92\x29\xC0\x34"
"\xCB\x62\xF9\x6D\x04\x9B\x0F\xA6"
"\x3D\xD4\x48\xDF\x76\x0D\x81\x18"
"\xAF\x23\xBA\x51\xE8\x5C\xF3\x8A"
"\x21\x95\x2C\xC3\x37\xCE\x65\xFC"
"\x70\x07\x9E\x12\xA9\x40\xD7\x4B"
"\xE2\x79\x10\x84\x1B\xB2\x26\xBD"
"\x54\xEB\x5F\xF6\x8D\x01\x98\x2F"
"\xC6\x3A\xD1\x68\xFF\x73\x0A\xA1"
"\x15\xAC\x43\xDA\x4E\xE5\x7C\x13"
"\x87\x1E\xB5\x29\xC0\x57\xEE\x62"
"\xF9\x90\x04\x9B\x32\xC9\x3D\xD4"
"\x6B\x02\x76\x0D\xA4\x18\xAF\x46"
"\xDD\x51\xE8\x7F\x16\x8A\x21\xB8"
"\x2C\xC3\x5A\xF1\x65\xFC\x93\x07"
"\x9E\x35\xCC\x40\xD7\x6E\x05\x79"
"\x10\xA7\x1B\xB2\x49\xE0\x54\xEB"
"\x82\x19\x8D\x24\xBB\x2F\xC6\x5D"
"\xF4\x68\xFF\x96\x0A\xA1\x38\xCF"
"\x43\xDA\x71\x08\x7C\x13\xAA\x1E"
"\xB5\x4C\xE3\x57\xEE\x85\x1C\x90"
"\x27\xBE\x32\xC9\x60\xF7\x6B\x02"
"\x99\x0D\xA4\x3B\xD2\x46\xDD\x74"
"\x0B\x7F\x16\xAD\x21\xB8\x4F\xE6"
"\x5A\xF1\x88\x1F\x93\x2A\xC1\x35"
"\xCC\x63\xFA\x6E\x05\x9C\x10\xA7"
"\x3E\xD5\x49\xE0\x77\x0E\x82\x19"
"\xB0\x24\xBB\x52\xE9\x5D\xF4\x8B"
"\x22\x96\x2D\xC4\x38\xCF\x66\xFD"
"\x71\x08\x9F\x13\xAA\x41\xD8\x4C"
"\xE3\x7A\x11\x85\x1C\xB3\x27\xBE"
"\x55\xEC\x60\xF7\x8E\x02\x99\x30"
"\xC7\x3B\xD2\x69\x00\x74\x0B\xA2"
"\x16\xAD\x44\xDB\x4F\xE6\x7D\x14"
"\x88\x1F\xB6\x2A\xC1\x58\xEF\x63"
"\xFA\x91\x05\x9C\x33\xCA\x3E\xD5"
"\x6C\x03\x77\x0E\xA5\x19\xB0\x47"
"\xDE\x52\xE9\x80\x17\x8B\x22\xB9"
"\x2D\xC4\x5B\xF2\x66\xFD\x94\x08"
"\x9F\x36\xCD\x41\xD8\x6F\x06\x7A"
"\x11\xA8\x1C\xB3\x4A\xE1\x55\xEC"
"\x83\x1A\x8E\x25\xBC\x30\xC7\x5E"
"\xF5\x69\x00\x97\x0B\xA2\x39\xD0"
"\x44\xDB\x72\x09\x7D\x14\xAB\x1F"
"\xB6\x4D\xE4\x58\xEF\x86\x1D\x91"
"\x28\xBF\x33\xCA\x61\xF8\x6C\x03"
"\x9A\x0E\xA5\x3C\xD3\x47\xDE\x75"
"\x0C\x80\x17\xAE\x22\xB9\x50\xE7"
"\x5B\xF2\x89\x20\x94\x2B\xC2\x36"
"\xCD\x64\xFB\x6F\x06\x9D\x11\xA8"
"\x3F\xD6\x4A\xE1\x78\x0F\x83\x1A"
"\xB1\x25\xBC\x53\xEA\x5E\xF5\x8C"
"\x00\x97\x2E\xC5\x39\xD0\x67\xFE"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x72\x09\xA0\x14\xAB\x42\xD9\x4D",
.ctext = "\xED\xCD\xDB\xB8\x68\xCE\xBD\xEA"
"\x9D\x9D\xCD\x9F\x4F\xFC\x4D\xB7"
"\xA5\xFF\x6F\x43\x0F\xBA\x32\x04"
"\xB3\xC2\xB9\x03\xAA\x91\x56\x29"
"\x0D\xD0\xFD\xC4\x65\xA5\x69\xB9"
"\xF1\xF6\xB1\xA5\xB2\x75\x4F\x8A"
"\x8D\x7D\x1B\x9B\xC7\x68\x72\xF8"
"\x01\x9B\x17\x0A\x29\xE7\x61\x28"
"\x7F\xA7\x50\xCA\x20\x2C\x96\x3B"
"\x6E\x5C\x5D\x3F\xB5\x7F\xF3\x2B"
"\x04\xEF\x9D\xD4\xCE\x41\x28\x8E"
"\x83\x54\xAE\x7C\x82\x46\x10\xC9"
"\xC4\x8A\x1E\x1F\x4C\xA9\xFC\xEC"
"\x3C\x8C\x30\xFC\x59\xD2\x54\xC4"
"\x6F\x50\xC6\xCA\x8C\x14\x5B\x9C"
"\x18\x56\x5B\xF8\x33\x0E\x4A\xDB"
"\xEC\xB5\x6E\x5B\x31\xC4\x0E\x98"
"\x9F\x32\xBA\xA2\x18\xCF\x55\x43"
"\xFE\x80\x8F\x60\xCF\x05\x30\x9B"
"\x70\x50\x1E\x9C\x08\x87\xE6\x20"
"\xD2\xF3\x27\xF8\x2A\x8D\x12\xB2"
"\xBC\x5F\xFE\x52\x52\xF6\x7F\xB6"
"\xB8\x30\x86\x3B\x0F\x94\x1E\x79"
"\x13\x94\x35\xA2\xB1\x35\x5B\x05"
"\x2A\x98\x6B\x96\x4C\xB1\x20\xBE"
"\xB6\x14\xC2\x06\xBF\xFD\x5F\x2A"
"\xF5\x33\xC8\x19\x45\x14\x44\x5D"
"\xFE\x94\x7B\xBB\x63\x13\x57\xC3"
"\x2A\x8F\x6C\x11\x2A\x07\xA7\x6A"
"\xBF\x20\xD3\x99\xC6\x00\x0B\xBF"
"\x83\x46\x25\x3A\xB0\xF6\xC5\xC8"
"\x00\xCA\xE5\x28\x4A\x7C\x95\x9C"
"\x7B\x43\xAB\xF9\xE4\xF8\x74\xAB"
"\xA7\xB8\x9C\x0F\x53\x7B\xB6\x74"
"\x60\x64\x0D\x1C\x80\xD1\x20\x9E"
"\xDC\x14\x27\x9B\xFC\xBD\x5C\x96"
"\xD2\x51\xDC\x96\xEE\xE5\xEA\x2B"
"\x02\x7C\xAA\x3C\xDC\x9D\x7B\x01"
"\x20\xC3\xE1\x0B\xDD\xAB\xF3\x1E"
"\x19\xA8\x84\x29\x5F\xCC\xC3\x5B"
"\xE4\x33\x59\xDC\x12\xEB\x2B\x4D"
"\x5B\x55\x23\xB7\x40\x31\xDE\xEE"
"\x18\xC9\x3C\x4D\xBC\xED\xE0\x42"
"\xAD\xDE\xA0\xA3\xC3\xFE\x44\xD3"
"\xE1\x9A\xDA\xAB\x32\xFC\x1A\xBF"
"\x63\xA9\xF0\x6A\x08\x46\xBD\x48"
"\x83\x06\xAB\x82\x99\x01\x16\x1A"
"\x03\x36\xC5\x59\x6B\xB8\x8C\x9F"
"\xC6\x51\x3D\xE5\x7F\xBF\xAB\xBC"
"\xC9\xA1\x88\x34\x5F\xA9\x7C\x3B"
"\x9F\x1B\x98\x2B\x4F\xFB\x9B\xF0"
"\xCD\xB6\x45\xB2\x29\x2E\x34\x23"
"\xA9\x97\xC0\x22\x8C\x42\x9B\x5F"
"\x40\xC8\xD7\x3D\x82\x9A\x6F\xAA"
"\x74\x83\x29\x05\xE8\xC4\x4D\x01"
"\xB5\xE5\x84\x3F\x7F\xD3\xE0\x99"
"\xDA\xE7\x6F\x30\xFD\xAA\x92\x30"
"\xA5\x46\x8B\xA2\xE6\x58\x62\x7C"
"\x2C\x35\x1B\x38\x85\x7D\xE8\xF3"
"\x87\x4F\xDA\xD8\x5F\xFC\xB6\x44"
"\xD0\xE3\x9B\x8B\xBF\xD6\xB8\xC4"
"\x73\xAE\x1D\x8B\x5B\x74\x8B\xCB"
"\xA4\xAD\xCF\x5D\xD4\x58\xC9\xCD"
"\xF7\x90\x68\xCF\xC9\x11\x52\x3E"
"\xE8\xA1\xA3\x78\x8B\xD0\xAC\x0A"
"\xD4\xC9\xA3\xA5\x55\x30\xC8\x3E"
"\xED\x28\x39\xE9\x63\xED\x41\x70"
"\x51\xE3\xC4\xA0\xFC\xD5\x43\xCB"
"\x4D\x65\xC8\xFD\x3A\x91\x8F\x60"
"\x8A\xA6\x6D\x9D\x3E\x01\x23\x4B"
"\x50\x47\xC9\xDC\x9B\xDE\x37\xC5"
"\xBF\x67\xB1\x6B\x78\x38\xD5\x7E"
"\xB6\xFF\x67\x83\x3B\x6E\xBE\x23"
"\x45\xFA\x1D\x69\x44\xFD\xC6\xB9"
"\xD0\x4A\x92\xD1\xBE\xF6\x4A\xB7"
"\xCA\xA8\xA2\x9E\x13\x87\x57\x92"
"\x64\x7C\x85\x0B\xB3\x29\x37\xD8"
"\xE6\xAA\xAF\xC4\x03\x67\xA3\xBF"
"\x2E\x45\x83\xB6\xD8\x54\x00\x89"
"\xF6\xBC\x3A\x7A\x88\x58\x51\xED"
"\xF4\x4E\x01\xA5\xC3\x2E\xD9\x42"
"\xBD\x6E\x0D\x0B\x21\xB0\x1A\xCC"
"\xA4\xD3\x3F\xDC\x9B\x81\xD8\xF1"
"\xEA\x7A\x6A\xB7\x07\xC9\x6D\x91"
"\x6D\x3A\xF5\x5F\xA6\xFF\x87\x1E"
"\x3F\xDD\xC0\x72\xEA\xAC\x08\x15"
"\x21\xE6\xC6\xB6\x0D\xD8\x51\x86"
"\x2A\x03\x73\xF7\x29\xD4\xC4\xE4"
"\x7F\x95\x10\xF7\xAB\x3F\x92\x23"
"\xD3\xCE\x9C\x2E\x46\x3B\x63\x43"
"\xBB\xC2\x82\x7A\x83\xD5\x55\xE2"
"\xE7\x9B\x2F\x92\xAF\xFD\x81\x56"
"\x79\xFD\x3E\xF9\x46\xE0\x25\xD4"
"\x38\xDE\xBC\x2C\xC4\x7A\x2A\x8F"
"\x94\x4F\xD0\xAD\x9B\x37\x18\xD4"
"\x0E\x4D\x0F\x02\x3A\xDC\x5A\xA2"
"\x39\x25\x55\x20\x5A\xA6\x02\x9F"
"\xE6\x77\x21\x77\xE5\x4B\x7B\x0B"
"\x30\xF8\x5F\x33\x0F\x49\xCD\xFF"
"\xF2\xE4\x35\xF9\xF0\x63\xC3\x7E"
"\xF1\xA6\x73\xB4\xDF\xE7\xBB\x78"
"\xFF\x21\xA9\xF3\xF3\xCF\x5D\xBA"
"\xED\x87\x98\xAC\xFE\x48\x97\x6D"
"\xA6\x7F\x69\x31\xB1\xC4\xFF\x14"
"\xC6\x76\xD4\x10\xDD\xF6\x49\x2C"
"\x9C\xC8\x6D\x76\xC0\x8F\x5F\x55"
"\x2F\x3C\x8A\x30\xAA\xC3\x16\x55"
"\xC6\xFC\x8D\x8B\xB9\xE5\x80\x6C"
"\xC8\x7E\xBD\x65\x58\x36\xD5\xBC"
"\xF0\x33\x52\x29\x70\xF9\x5C\xE9"
"\xAC\x1F\xB5\x73\x56\x66\x54\xAF"
"\x1B\x8F\x7D\xED\xAB\x03\xCE\xE3"
"\xAE\x47\xB6\x69\x86\xE9\x01\x31"
"\x83\x18\x3D\xF4\x74\x7B\xF9\x42"
"\x4C\xFD\x75\x4A\x6D\xF0\x03\xA6"
"\x2B\x20\x63\xDA\x49\x65\x5E\x8B"
"\xC0\x19\xE3\x8D\xD9\xF3\xB0\x34"
"\xD3\x52\xFC\x68\x00\x43\x1B\x37"
"\x31\x93\x51\x1C\x63\x97\x70\xB0"
"\x99\x78\x83\x13\xFD\xCF\x53\x81"
"\x36\x46\xB5\x42\x52\x2F\x32\xEB"
"\x4A\x3D\xF1\x8F\x1C\x54\x2E\xFC"
"\x41\x75\x5A\x8C\x8E\x6F\xE7\x1A"
"\xAE\xEF\x3E\x82\x12\x0B\x74\x72"
"\xF8\xB2\xAA\x7A\xD6\xFF\xFA\x55"
"\x33\x1A\xBB\xD3\xA2\x7E\x97\x66",
.len = 1008,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.np = 3,
.tap = { 1008 - 20, 4, 16 },
},
};
static const struct cipher_testvec camellia_cbc_tv_template[] = {
{
.key = "\x06\xa9\x21\x40\x36\xb8\xa1\x5b"
"\x51\x2e\x03\xd5\x34\x12\x00\x06",
.klen = 16,
.iv = "\x3d\xaf\xba\x42\x9d\x9e\xb4\x30"
"\xb4\x22\xda\x80\x2c\x9f\xac\x41",
.ptext = "Single block msg",
.ctext = "\xea\x32\x12\x76\x3b\x50\x10\xe7"
"\x18\xf6\xfd\x5d\xf6\x8f\x13\x51",
.len = 16,
}, {
.key = "\xc2\x86\x69\x6d\x88\x7c\x9a\xa0"
"\x61\x1b\xbb\x3e\x20\x25\xa4\x5a",
.klen = 16,
.iv = "\x56\x2e\x17\x99\x6d\x09\x3d\x28"
"\xdd\xb3\xba\x69\x5a\x2e\x6f\x58",
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.ctext = "\xa5\xdf\x6e\x50\xda\x70\x6c\x01"
"\x4a\xab\xf3\xf2\xd6\xfc\x6c\xfd"
"\x19\xb4\x3e\x57\x1c\x02\x5e\xa0"
"\x15\x78\xe0\x5e\xf2\xcb\x87\x16",
.len = 32,
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2 Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-13 04:46:35 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06"
"\x9D\x34\xCB\x3F\xD6\x6D\x04\x78"
"\x0F\xA6\x1A\xB1\x48\xDF\x53\xEA"
"\x81\x18\x8C\x23\xBA\x2E\xC5\x5C"
"\xF3\x67\xFE\x95\x09\xA0\x37\xCE"
"\x42\xD9\x70\x07\x7B\x12\xA9\x1D"
"\xB4\x4B\xE2\x56\xED\x84\x1B\x8F"
"\x26\xBD\x31\xC8\x5F\xF6\x6A\x01"
"\x98\x0C\xA3\x3A\xD1\x45\xDC\x73"
"\x0A\x7E\x15\xAC\x20\xB7\x4E\xE5"
"\x59\xF0\x87\x1E\x92\x29\xC0\x34"
"\xCB\x62\xF9\x6D\x04\x9B\x0F\xA6"
"\x3D\xD4\x48\xDF\x76\x0D\x81\x18"
"\xAF\x23\xBA\x51\xE8\x5C\xF3\x8A"
"\x21\x95\x2C\xC3\x37\xCE\x65\xFC"
"\x70\x07\x9E\x12\xA9\x40\xD7\x4B"
"\xE2\x79\x10\x84\x1B\xB2\x26\xBD"
"\x54\xEB\x5F\xF6\x8D\x01\x98\x2F"
"\xC6\x3A\xD1\x68\xFF\x73\x0A\xA1"
"\x15\xAC\x43\xDA\x4E\xE5\x7C\x13"
"\x87\x1E\xB5\x29\xC0\x57\xEE\x62"
"\xF9\x90\x04\x9B\x32\xC9\x3D\xD4"
"\x6B\x02\x76\x0D\xA4\x18\xAF\x46"
"\xDD\x51\xE8\x7F\x16\x8A\x21\xB8"
"\x2C\xC3\x5A\xF1\x65\xFC\x93\x07"
"\x9E\x35\xCC\x40\xD7\x6E\x05\x79"
"\x10\xA7\x1B\xB2\x49\xE0\x54\xEB"
"\x82\x19\x8D\x24\xBB\x2F\xC6\x5D"
"\xF4\x68\xFF\x96\x0A\xA1\x38\xCF"
"\x43\xDA\x71\x08\x7C\x13\xAA\x1E"
"\xB5\x4C\xE3\x57\xEE\x85\x1C\x90"
"\x27\xBE\x32\xC9\x60\xF7\x6B\x02"
"\x99\x0D\xA4\x3B\xD2\x46\xDD\x74"
"\x0B\x7F\x16\xAD\x21\xB8\x4F\xE6"
"\x5A\xF1\x88\x1F\x93\x2A\xC1\x35"
"\xCC\x63\xFA\x6E\x05\x9C\x10\xA7"
"\x3E\xD5\x49\xE0\x77\x0E\x82\x19"
"\xB0\x24\xBB\x52\xE9\x5D\xF4\x8B"
"\x22\x96\x2D\xC4\x38\xCF\x66\xFD"
"\x71\x08\x9F\x13\xAA\x41\xD8\x4C"
"\xE3\x7A\x11\x85\x1C\xB3\x27\xBE"
"\x55\xEC\x60\xF7\x8E\x02\x99\x30"
"\xC7\x3B\xD2\x69\x00\x74\x0B\xA2"
"\x16\xAD\x44\xDB\x4F\xE6\x7D\x14"
"\x88\x1F\xB6\x2A\xC1\x58\xEF\x63"
"\xFA\x91\x05\x9C\x33\xCA\x3E\xD5"
"\x6C\x03\x77\x0E\xA5\x19\xB0\x47"
"\xDE\x52\xE9\x80\x17\x8B\x22\xB9"
"\x2D\xC4\x5B\xF2\x66\xFD\x94\x08"
"\x9F\x36\xCD\x41\xD8\x6F\x06\x7A"
"\x11\xA8\x1C\xB3\x4A\xE1\x55\xEC"
"\x83\x1A\x8E\x25\xBC\x30\xC7\x5E"
"\xF5\x69\x00\x97\x0B\xA2\x39\xD0"
"\x44\xDB\x72\x09\x7D\x14\xAB\x1F"
"\xB6\x4D\xE4\x58\xEF\x86\x1D\x91"
"\x28\xBF\x33\xCA\x61\xF8\x6C\x03"
"\x9A\x0E\xA5\x3C\xD3\x47\xDE\x75"
"\x0C\x80\x17\xAE\x22\xB9\x50\xE7"
"\x5B\xF2\x89\x20\x94\x2B\xC2\x36"
"\xCD\x64\xFB\x6F\x06\x9D\x11\xA8"
"\x3F\xD6\x4A\xE1\x78\x0F\x83\x1A"
"\xB1\x25\xBC\x53\xEA\x5E\xF5\x8C"
"\x00\x97\x2E\xC5\x39\xD0\x67\xFE"
"\x72\x09\xA0\x14\xAB\x42\xD9\x4D",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xCD\x3E\x2A\x3B\x3E\x94\xC5\x77"
"\xBA\xBB\x5B\xB1\xDE\x7B\xA4\x40"
"\x88\x39\xE3\xFD\x94\x4B\x25\x58"
"\xE1\x4B\xC4\x18\x7A\xFD\x17\x2B"
"\xB9\xF9\xC2\x27\x6A\xB6\x31\x27"
"\xA6\xAD\xEF\xE5\x5D\xE4\x02\x01"
"\x56\x2E\x10\xC2\x2C\xFF\xC6\x83"
"\xB5\xDC\x4F\x63\xAD\x0E\x63\x5E"
"\x56\xC8\x18\x3D\x79\x86\x97\xEF"
"\x57\x0E\x63\xA1\xC1\x41\x48\xB8"
"\x98\xB7\x51\x6D\x18\xF6\x19\x82"
"\x37\x49\x88\xA4\xEF\x91\x21\x47"
"\x03\x28\xEA\x42\xF4\xFB\x7A\x58"
"\x28\x90\x77\x46\xD8\xD2\x35\x16"
"\x44\xA9\x9E\x49\x52\x2A\xE4\x16"
"\x5D\xF7\x65\xEB\x0F\xC9\x29\xE6"
"\xCF\x76\x91\x89\x8A\x94\x39\xFA"
"\x6B\x5F\x63\x53\x74\x43\x91\xF5"
"\x3F\xBC\x88\x53\xB2\x1A\x02\x3F"
"\x9D\x32\x84\xEB\x56\x28\xD6\x06"
"\xD5\xB2\x20\xA9\xFC\xC3\x76\x62"
"\x32\xCC\x86\xC8\x36\x67\x5E\x7E"
"\xA4\xAA\x15\x63\x6B\xA9\x86\xAF"
"\x1A\x52\x82\x36\x5F\xF4\x3F\x7A"
"\x9B\x78\x62\x3B\x02\x28\x60\xB3"
"\xBA\x82\xB1\xDD\xC9\x60\x8F\x47"
"\xF1\x6B\xFE\xE5\x39\x34\xA0\x28"
"\xA4\xB3\xC9\x7E\xED\x28\x8D\x70"
"\xB2\x1D\xFD\xC6\x00\xCF\x1A\x94"
"\x28\xF8\xC1\x34\xB7\x58\xA5\x6C"
"\x1A\x9D\xE4\xE4\xF6\xB9\xB4\xB0"
"\x5D\x51\x54\x9A\x53\xA0\xF9\x32"
"\xBD\x31\x54\x14\x7B\x33\xEE\x17"
"\xD3\xC7\x1F\x48\xBF\x0B\x22\xA2"
"\x7D\x0C\xDF\xD0\x2E\x98\xFA\xD2"
"\xFA\xCF\x24\x1D\x99\x9B\xD0\x7E"
"\xF4\x4F\x88\xFF\x45\x99\x4A\xF4"
"\xF2\x0A\x5B\x3B\x21\xAB\x92\xAE"
"\x40\x78\x91\x95\xC4\x2F\xA3\xE8"
"\x18\xC7\x07\xA6\xC8\xC0\x66\x33"
"\x35\xC0\xB4\xA0\xF8\xEE\x1E\xF3"
"\x40\xF5\x40\x54\xF1\x84\x8C\xEA"
"\x27\x38\x1F\xF8\x77\xC7\xDF\xD8"
"\x1D\xE2\xD9\x59\x40\x4F\x59\xD4"
"\xF8\x17\x99\x8D\x58\x2D\x72\x44"
"\x9D\x1D\x91\x64\xD6\x3F\x0A\x82"
"\xC7\x57\x3D\xEF\xD3\x41\xFA\xA7"
"\x68\xA3\xB8\xA5\x93\x74\x2E\x85"
"\x4C\x9D\x69\x59\xCE\x15\xAE\xBF"
"\x9C\x8F\x14\x64\x5D\x7F\xCF\x0B"
"\xCE\x43\x5D\x28\xC0\x2F\xFB\x18"
"\x79\x9A\xFC\x43\x16\x7C\x6B\x7B"
"\x38\xB8\x48\x36\x66\x4E\x20\x43"
"\xBA\x76\x13\x9A\xC3\xF2\xEB\x52"
"\xD7\xDC\xB2\x67\x63\x14\x25\xCD"
"\xB1\x13\x4B\xDE\x8C\x59\x21\x84"
"\x81\x8D\x97\x23\x45\x33\x7C\xF3"
"\xC5\xBC\x79\x95\xAA\x84\x68\x31"
"\x2D\x1A\x68\xFE\xEC\x92\x94\xDA"
"\x94\x2A\x6F\xD6\xFE\xE5\x76\x97"
"\xF4\x6E\xEE\xCB\x2B\x95\x4E\x36"
"\x5F\x74\x8C\x86\x5B\x71\xD0\x20"
"\x78\x1A\x7F\x18\x8C\xD9\xCD\xF5"
"\x21\x41\x56\x72\x13\xE1\x86\x07"
"\x07\x26\xF3\x4F\x7B\xEA\xB5\x18"
"\xFE\x94\x2D\x9F\xE0\x72\x18\x65"
"\xB2\xA5\x63\x48\xB4\x13\x22\xF7"
"\x25\xF1\x80\xA8\x7F\x54\x86\x7B"
"\x39\xAE\x95\x0C\x09\x32\x22\x2D"
"\x4D\x73\x39\x0C\x09\x2C\x7C\x10"
"\xD0\x4B\x53\xF6\x90\xC5\x99\x2F"
"\x15\xE1\x7F\xC6\xC5\x7A\x52\x14"
"\x65\xEE\x93\x54\xD0\x66\x15\x3C"
"\x4C\x68\xFD\x64\x0F\xF9\x10\x39"
"\x46\x7A\xDD\x97\x20\xEE\xC7\xD2"
"\x98\x4A\xB6\xE6\xF5\xA8\x1F\x4F"
"\xDB\xAB\x6D\xD5\x9B\x34\x16\x97"
"\x2F\x64\xE5\x37\xEF\x0E\xA1\xE9"
"\xBE\x31\x31\x96\x8B\x40\x18\x75"
"\x11\x75\x14\x32\xA5\x2D\x1B\x6B"
"\xDB\x59\xEB\xFA\x3D\x8E\x7C\xC4"
"\xDE\x68\xC8\x9F\xC9\x99\xE3\xC6"
"\x71\xB0\x12\x57\x89\x0D\xC0\x2B"
"\x9F\x12\x6A\x04\x67\xF1\x95\x31"
"\x59\xFD\x84\x95\x2C\x9C\x5B\xEC"
"\x09\xB0\x43\x96\x4A\x64\x80\x40"
"\xB9\x72\x19\xDD\x70\x42\xFA\xB1"
"\x4A\x2C\x0C\x0A\x60\x6E\xE3\x7C"
"\x37\x5A\xBE\xA4\x62\xCF\x29\xAB"
"\x7F\x4D\xA6\xB3\xE2\xB6\x64\xC6"
"\x33\x0B\xF3\xD5\x01\x38\x74\xA4"
"\x67\x1E\x75\x68\xC3\xAD\x76\xE9"
"\xE9\xBC\xF0\xEB\xD8\xFD\x31\x8A"
"\x5F\xC9\x18\x94\x4B\x86\x66\xFC"
"\xBD\x0B\x3D\xB3\x9F\xFA\x1F\xD9"
"\x78\xC4\xE3\x24\x1C\x67\xA2\xF8"
"\x43\xBC\x76\x75\xBF\x6C\x05\xB3"
"\x32\xE8\x7C\x80\xDB\xC7\xB6\x61"
"\x1A\x3E\x2B\xA7\x25\xED\x8F\xA0"
"\x00\x4B\xF8\x90\xCA\xD8\xFB\x12"
"\xAC\x1F\x18\xE9\xD2\x5E\xA2\x8E"
"\xE4\x84\x6B\x9D\xEB\x1E\x6B\xA3"
"\x7B\xDC\xCE\x15\x97\x27\xB2\x65"
"\xBC\x0E\x47\xAB\x55\x13\x53\xAB"
"\x0E\x34\x55\x02\x5F\x27\xC5\x89"
"\xDF\xC5\x70\xC4\xDD\x76\x82\xEE"
"\x68\xA6\x09\xB0\xE5\x5E\xF1\x0C"
"\xE3\xF3\x09\x9B\xFE\x65\x4B\xB8"
"\x30\xEC\xD5\x7C\x6A\xEC\x1D\xD2"
"\x93\xB7\xA1\x1A\x02\xD4\xC0\xD6"
"\x8D\x4D\x83\x9A\xED\x29\x4E\x14"
"\x86\xD5\x3C\x1A\xD5\xB9\x0A\x6A"
"\x72\x22\xD5\x92\x38\xF1\xA1\x86"
"\xB2\x41\x51\xCA\x4E\xAB\x8F\xD3"
"\x80\x56\xC3\xD7\x65\xE1\xB3\x86"
"\xCB\xCE\x98\xA1\xD4\x59\x1C\x06"
"\x01\xED\xF8\x29\x91\x19\x5C\x9A"
"\xEE\x28\x1B\x48\xD7\x32\xEF\x9F"
"\x6C\x2B\x66\x4E\x78\xD5\x8B\x72"
"\x80\xE7\x29\xDC\x23\x55\x98\x54"
"\xB1\xFF\x3E\x95\x56\xA8\x78\x78"
"\xEF\xC4\xA5\x11\x2D\x2B\xD8\x93"
"\x30\x6E\x7E\x51\xBB\x42\x5F\x03"
"\x43\x94\x23\x7E\xEE\xF0\xA5\x79"
"\x55\x01\xD4\x58\xB2\xF2\x85\x49"
"\x70\xC5\xB9\x0B\x3B\x7A\x6E\x6C",
.len = 1008,
.also_non_np = 1,
.np = 3,
.tap = { 1008 - 20, 4, 16 },
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec camellia_ctr_tv_template[] = {
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
{ /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7",
.ctext = "\xF3\x06\x3A\x84\xCD\xBA\x8E\x11"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xB7\x74\x6F\x5C\x97\xFB\x36\xFE"
"\xDE\x71\x58\xD4\x15\xD1\xC1\xA4"
"\xC9\x28\x74\xA6\x6B\xC7\x95\xA6"
"\x6C\x77\xF7\x2F\xDF\xC7\xBB\x85"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\x60\xFC\xE8\x94\xE8\xB5\x09\x2C"
"\x1E\x43\xEF\x6C\xE9\x98\xC5\xA0"
"\x7B\x13\xE5\x7F\xF8\x49\x9A\x8C"
"\xE6\x7B\x08\xC3\x32\x66\x55\x4E"
"\xA5\x44\x1D\x2C\x18\xC7\x29\x1F"
"\x61\x28\x4A\xE3\xCD\xE5\x47\xB2"
"\x82\x2F\x66\x83\x91\x51\xAE\xD7"
"\x1C\x91\x3C\x57\xE3\x1D\x5A\xC9"
"\xFD\xC5\x58\x58\xEF\xCC\x33\xC9"
"\x0F\xEA\x26\x32\xD1\x15\x19\x2D"
"\x25\xB4\x7F\xB0\xDF\xFB\x88\x60"
"\x4E\x4D\x06\x7D\xCC\x1F\xED\x3B"
"\x68\x84\xD5\xB3\x1B\xE7\xB9\xA1"
"\x68\x8B\x2C\x1A\x44\xDA\x63\xD3"
"\x29\xE9\x59\x32\x1F\x30\x1C\x43"
"\xEA\x3A\xA3\x6B\x54\x3C\xAA\x11"
"\xAD\x38\x20\xC9\xB9\x8A\x64\x66"
"\x5A\x07\x49\xDF\xA1\x9C\xF9\x76"
"\x36\x65\xB6\x81\x8F\x76\x09\xE5"
"\xEB\xD1\x29\xA4\xE4\xF4\x4C\xCD"
"\xAF\xFC\xB9\x16\xD9\xC3\x73\x6A"
"\x33\x12\xF8\x7E\xBC\xCC\x7D\x80"
"\xBF\x3C\x25\x06\x13\x84\xFA\x35"
"\xF7\x40\xFA\xA1\x44\x13\x70\xD8"
"\x01\xF9\x85\x15\x63\xEC\x7D\xB9"
"\x02\xD8\xBA\x41\x6C\x92\x68\x66"
"\x95\xDD\xD6\x42\xE7\xBB\xE1\xFD"
"\x28\x3E\x94\xB6\xBD\xA7\xBF\x47"
"\x58\x8D\xFF\x19\x30\x75\x0D\x48"
"\x94\xE9\xA6\xCD\xB3\x8E\x1E\xCD"
"\x59\xBC\x1A\xAC\x3C\x4F\xA9\xEB"
"\xF4\xA7\xE4\x75\x4A\x18\x40\xC9"
"\x1E\xEC\x06\x9C\x28\x4B\xF7\x2B"
"\xE2\xEF\xD6\x42\x2E\xBB\xFC\x0A"
"\x79\xA2\x99\x28\x93\x1B\x00\x57"
"\x35\x1E\x1A\x93\x90\xA4\x68\x95"
"\x5E\x57\x40\xD5\xA9\xAA\x19\x48"
"\xEC\xFF\x76\x77\xDC\x78\x89\x76"
"\xE5\x3B\x00\xEC\x58\x4D\xD1\xE3"
"\xC8\x6C\x2C\x45\x5E\x5F\xD9\x4E"
"\x71\xA5\x36\x6D\x03\xF1\xC7\xD5"
"\xF3\x63\xC0\xD8\xCB\x2B\xF1\xA8"
"\xB9\x2B\xE6\x0B\xB9\x65\x78\xA0"
"\xC4\x46\xE6\x9B\x8B\x43\x2D\xAB"
"\x70\xA6\xE0\x59\x1E\xAC\x9D\xE0"
"\x76\x44\x45\xF3\x24\x11\x57\x98"
"\x9A\x86\xB4\x12\x80\x28\x86\x20"
"\x23\x9D\x2D\xE9\x38\x32\xB1\xE1"
"\xCF\x0A\x23\x73\x7D\xC5\x80\x3D"
"\x9F\x6D\xA0\xD0\xEE\x93\x8A\x79"
"\x3A\xDD\x1D\xBB\x9E\x26\x5D\x01"
"\x44\xD0\xD4\x4E\xC3\xF1\xE4\x38"
"\x09\x62\x0A\x1A\x4E\xD2\x63\x0F"
"\x6E\x3E\xD2\xA4\x3A\xF4\xF3\xFF"
"\x7E\x42\xEC\xB6\x6F\x4D\x6B\x48"
"\xE6\xA6\x50\x80\x78\x9E\xF1\xB0"
"\x4D\xB2\x0D\x3D\xFC\x40\x25\x4D",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 496,
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xE2\x24\x89\xEE\x53\xB8\x1D\x5F"
"\xC4\x29\x8E\xF3\x35\x9A\xFF\x64",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06"
"\x9D\x34\xCB\x3F\xD6\x6D\x04\x78"
"\x0F\xA6\x1A\xB1\x48\xDF\x53\xEA"
"\x81\x18\x8C\x23\xBA\x2E\xC5\x5C"
"\xF3\x67\xFE\x95\x09\xA0\x37\xCE"
"\x42\xD9\x70\x07\x7B\x12\xA9\x1D"
"\xB4\x4B\xE2\x56\xED\x84\x1B\x8F"
"\x26\xBD\x31\xC8\x5F\xF6\x6A\x01"
"\x98\x0C\xA3\x3A\xD1\x45\xDC\x73"
"\x0A\x7E\x15\xAC\x20\xB7\x4E\xE5"
"\x59\xF0\x87\x1E\x92\x29\xC0\x34"
"\xCB\x62\xF9\x6D\x04\x9B\x0F\xA6"
"\x3D\xD4\x48\xDF\x76\x0D\x81\x18"
"\xAF\x23\xBA\x51\xE8\x5C\xF3\x8A"
"\x21\x95\x2C\xC3\x37\xCE\x65\xFC"
"\x70\x07\x9E\x12\xA9\x40\xD7\x4B"
"\xE2\x79\x10\x84\x1B\xB2\x26\xBD"
"\x54\xEB\x5F\xF6\x8D\x01\x98\x2F"
"\xC6\x3A\xD1\x68\xFF\x73\x0A\xA1"
"\x15\xAC\x43\xDA\x4E\xE5\x7C\x13"
"\x87\x1E\xB5\x29\xC0\x57\xEE\x62"
"\xF9\x90\x04\x9B\x32\xC9\x3D\xD4"
"\x6B\x02\x76\x0D\xA4\x18\xAF\x46"
"\xDD\x51\xE8\x7F\x16\x8A\x21\xB8"
"\x2C\xC3\x5A\xF1\x65\xFC\x93\x07"
"\x9E\x35\xCC\x40\xD7\x6E\x05\x79"
"\x10\xA7\x1B\xB2\x49\xE0\x54\xEB"
"\x82\x19\x8D\x24\xBB\x2F\xC6\x5D"
"\xF4\x68\xFF\x96\x0A\xA1\x38\xCF"
"\x43\xDA\x71\x08\x7C\x13\xAA\x1E"
"\xB5\x4C\xE3\x57\xEE\x85\x1C\x90"
"\x27\xBE\x32\xC9\x60\xF7\x6B\x02"
"\x99\x0D\xA4\x3B\xD2\x46\xDD\x74"
"\x0B\x7F\x16\xAD\x21\xB8\x4F\xE6"
"\x5A\xF1\x88\x1F\x93\x2A\xC1\x35"
"\xCC\x63\xFA\x6E\x05\x9C\x10\xA7"
"\x3E\xD5\x49\xE0\x77\x0E\x82\x19"
"\xB0\x24\xBB\x52\xE9\x5D\xF4\x8B"
"\x22\x96\x2D\xC4\x38\xCF\x66\xFD"
"\x71\x08\x9F\x13\xAA\x41\xD8\x4C"
"\xE3\x7A\x11\x85\x1C\xB3\x27\xBE"
"\x55\xEC\x60\xF7\x8E\x02\x99\x30"
"\xC7\x3B\xD2\x69\x00\x74\x0B\xA2"
"\x16\xAD\x44\xDB\x4F\xE6\x7D\x14"
"\x88\x1F\xB6\x2A\xC1\x58\xEF\x63"
"\xFA\x91\x05\x9C\x33\xCA\x3E\xD5"
"\x6C\x03\x77\x0E\xA5\x19\xB0\x47"
"\xDE\x52\xE9\x80\x17\x8B\x22\xB9"
"\x2D\xC4\x5B\xF2\x66\xFD\x94\x08"
"\x9F\x36\xCD\x41\xD8\x6F\x06\x7A"
"\x11\xA8\x1C\xB3\x4A\xE1\x55\xEC"
"\x83\x1A\x8E\x25\xBC\x30\xC7\x5E"
"\xF5\x69\x00\x97\x0B\xA2\x39\xD0"
"\x44\xDB\x72\x09\x7D\x14\xAB\x1F"
"\xB6\x4D\xE4\x58\xEF\x86\x1D\x91"
"\x28\xBF\x33\xCA\x61\xF8\x6C\x03"
"\x9A\x0E\xA5\x3C\xD3\x47\xDE\x75"
"\x0C\x80\x17\xAE\x22\xB9\x50\xE7"
"\x5B\xF2\x89\x20\x94\x2B\xC2\x36"
"\xCD\x64\xFB\x6F\x06\x9D\x11\xA8"
"\x3F\xD6\x4A\xE1\x78\x0F\x83\x1A"
"\xB1\x25\xBC\x53\xEA\x5E\xF5\x8C"
"\x00\x97\x2E\xC5\x39\xD0\x67\xFE"
"\x72\x09\xA0\x14\xAB\x42\xD9\x4D"
"\xE4\x7B\x12",
.ctext = "\xF3\x06\x3A\x84\xCD\xBA\x8E\x11"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xB7\x74\x6F\x5C\x97\xFB\x36\xFE"
"\xDE\x71\x58\xD4\x15\xD1\xC1\xA4"
"\xC9\x28\x74\xA6\x6B\xC7\x95\xA6"
"\x6C\x77\xF7\x2F\xDF\xC7\xBB\x85"
"\x60\xFC\xE8\x94\xE8\xB5\x09\x2C"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\x1E\x43\xEF\x6C\xE9\x98\xC5\xA0"
"\x7B\x13\xE5\x7F\xF8\x49\x9A\x8C"
"\xE6\x7B\x08\xC3\x32\x66\x55\x4E"
"\xA5\x44\x1D\x2C\x18\xC7\x29\x1F"
"\x61\x28\x4A\xE3\xCD\xE5\x47\xB2"
"\x82\x2F\x66\x83\x91\x51\xAE\xD7"
"\x1C\x91\x3C\x57\xE3\x1D\x5A\xC9"
"\xFD\xC5\x58\x58\xEF\xCC\x33\xC9"
"\x0F\xEA\x26\x32\xD1\x15\x19\x2D"
"\x25\xB4\x7F\xB0\xDF\xFB\x88\x60"
"\x4E\x4D\x06\x7D\xCC\x1F\xED\x3B"
"\x68\x84\xD5\xB3\x1B\xE7\xB9\xA1"
"\x68\x8B\x2C\x1A\x44\xDA\x63\xD3"
"\x29\xE9\x59\x32\x1F\x30\x1C\x43"
"\xEA\x3A\xA3\x6B\x54\x3C\xAA\x11"
"\xAD\x38\x20\xC9\xB9\x8A\x64\x66"
"\x5A\x07\x49\xDF\xA1\x9C\xF9\x76"
"\x36\x65\xB6\x81\x8F\x76\x09\xE5"
"\xEB\xD1\x29\xA4\xE4\xF4\x4C\xCD"
"\xAF\xFC\xB9\x16\xD9\xC3\x73\x6A"
"\x33\x12\xF8\x7E\xBC\xCC\x7D\x80"
"\xBF\x3C\x25\x06\x13\x84\xFA\x35"
"\xF7\x40\xFA\xA1\x44\x13\x70\xD8"
"\x01\xF9\x85\x15\x63\xEC\x7D\xB9"
"\x02\xD8\xBA\x41\x6C\x92\x68\x66"
"\x95\xDD\xD6\x42\xE7\xBB\xE1\xFD"
"\x28\x3E\x94\xB6\xBD\xA7\xBF\x47"
"\x58\x8D\xFF\x19\x30\x75\x0D\x48"
"\x94\xE9\xA6\xCD\xB3\x8E\x1E\xCD"
"\x59\xBC\x1A\xAC\x3C\x4F\xA9\xEB"
"\xF4\xA7\xE4\x75\x4A\x18\x40\xC9"
"\x1E\xEC\x06\x9C\x28\x4B\xF7\x2B"
"\xE2\xEF\xD6\x42\x2E\xBB\xFC\x0A"
"\x79\xA2\x99\x28\x93\x1B\x00\x57"
"\x35\x1E\x1A\x93\x90\xA4\x68\x95"
"\x5E\x57\x40\xD5\xA9\xAA\x19\x48"
"\xEC\xFF\x76\x77\xDC\x78\x89\x76"
"\xE5\x3B\x00\xEC\x58\x4D\xD1\xE3"
"\xC8\x6C\x2C\x45\x5E\x5F\xD9\x4E"
"\x71\xA5\x36\x6D\x03\xF1\xC7\xD5"
"\xF3\x63\xC0\xD8\xCB\x2B\xF1\xA8"
"\xB9\x2B\xE6\x0B\xB9\x65\x78\xA0"
"\xC4\x46\xE6\x9B\x8B\x43\x2D\xAB"
"\x70\xA6\xE0\x59\x1E\xAC\x9D\xE0"
"\x76\x44\x45\xF3\x24\x11\x57\x98"
"\x9A\x86\xB4\x12\x80\x28\x86\x20"
"\x23\x9D\x2D\xE9\x38\x32\xB1\xE1"
"\xCF\x0A\x23\x73\x7D\xC5\x80\x3D"
"\x9F\x6D\xA0\xD0\xEE\x93\x8A\x79"
"\x3A\xDD\x1D\xBB\x9E\x26\x5D\x01"
"\x44\xD0\xD4\x4E\xC3\xF1\xE4\x38"
"\x09\x62\x0A\x1A\x4E\xD2\x63\x0F"
"\x6E\x3E\xD2\xA4\x3A\xF4\xF3\xFF"
"\x7E\x42\xEC\xB6\x6F\x4D\x6B\x48"
"\xE6\xA6\x50\x80\x78\x9E\xF1\xB0"
"\x4D\xB2\x0D\x3D\xFC\x40\x25\x4D"
crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2 Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-13 04:46:35 -06:00
"\x93\x11\x1C\xE9\xD2\x9F\x6E\x90"
"\xE5\x41\x4A\xE2\x3C\x45\x29\x35"
"\xEC\xD6\x47\x50\xCB\x7B\xA2\x32"
"\xF7\x8B\x62\xF1\xE3\x9A\xFE\xC7"
"\x1D\x8C\x02\x72\x68\x09\xE9\xB6"
"\x4A\x80\xE6\xB1\x56\xDF\x90\xD4"
"\x93\x74\xA4\xCE\x20\x23\xBF\x48"
"\xA5\xDE\x1B\xFA\x40\x69\x31\x98"
"\x62\x6E\xA5\xC7\xBF\x0C\x62\xE5"
"\x6D\xE1\x93\xF1\x83\x10\x1C\xCA"
"\xF6\x5C\x19\xF8\x90\x78\xCB\xE4"
"\x0B\x3A\xB5\xF8\x43\x86\xD3\x3F"
"\xBA\x83\x34\x3C\x42\xCC\x7D\x28"
"\x29\x63\x4F\xD8\x02\x17\xC5\x07"
"\x2C\xA4\xAC\x79\xCB\xC3\xA9\x09"
"\x81\x45\x18\xED\xE4\xCB\x42\x3B"
"\x87\x2D\x23\xDC\xC5\xBA\x45\xBD"
"\x92\xE5\x02\x97\x96\xCE\xAD\xEC"
"\xBA\xD8\x76\xF8\xCA\xC1\x31\xEC"
"\x1E\x4F\x3F\x83\xF8\x33\xE8\x6E"
"\xCC\xF8\x5F\xDD\x65\x50\x99\x69"
"\xAF\x48\xCE\xA5\xBA\xB6\x14\x9F"
"\x05\x93\xB2\xE6\x59\xC8\x28\xFE"
"\x8F\x37\xF9\x64\xB9\xA5\x56\x8F"
"\xF1\x1B\x90\xEF\xAE\xEB\xFC\x09"
"\x11\x7A\xF2\x19\x0A\x0A\x9A\x3C"
"\xE2\x5E\x29\xFA\x31\x9B\xC1\x74"
"\x1E\x10\x3E\x07\xA9\x31\x6D\xF8"
"\x81\xF5\xD5\x8A\x04\x23\x51\xAC"
"\xA2\xE2\x63\xFD\x27\x1F\x79\x5B"
"\x1F\xE8\xDA\x11\x49\x4D\x1C\xBA"
"\x54\xCC\x0F\xBA\x92\x69\xE5\xCB"
"\x41\x1A\x67\xA6\x40\x82\x70\x8C"
"\x19\x79\x08\xA4\x51\x20\x7D\xC9"
"\x12\x27\xAE\x20\x0D\x2C\xA1\x6D"
"\xF4\x55\xD4\xE7\xE6\xD4\x28\x08"
"\x00\x70\x12\x56\x56\x50\xAD\x14"
"\x5C\x3E\xA2\xD1\x36\x3F\x36\x48"
"\xED\xB1\x57\x3E\x5D\x15\xF6\x1E"
"\x53\xE9\xA4\x3E\xED\x7D\xCF\x7D"
"\x29\xAF\xF3\x1E\x51\xA8\x9F\x85"
"\x8B\xF0\xBB\xCE\xCC\x39\xC3\x64"
"\x4B\xF2\xAD\x70\x19\xD4\x44\x8F"
"\x91\x76\xE8\x15\x66\x34\x9F\xF6"
"\x0F\x15\xA4\xA8\x24\xF8\x58\xB1"
"\x38\x46\x47\xC7\x9B\xCA\xE9\x42"
"\x44\xAA\xE6\xB5\x9C\x91\xA4\xD3"
"\x16\xA0\xED\x42\xBE\xB5\x06\x19"
"\xBE\x67\xE8\xBC\x22\x32\xA4\x1E"
"\x93\xEB\xBE\xE9\xE1\x93\xE5\x31"
"\x3A\xA2\x75\xDF\xE3\x6B\xE7\xCC"
"\xB4\x70\x20\xE0\x6D\x82\x7C\xC8"
"\x94\x5C\x5E\x37\x18\xAD\xED\x8B"
"\x44\x86\xCA\x5E\x07\xB7\x70\x8D"
"\x40\x48\x19\x73\x7C\x78\x64\x0B"
"\xDB\x01\xCA\xAE\x63\x19\xE9\xD1"
"\x6B\x2C\x84\x10\x45\x42\x2E\xC3"
"\xDF\x7F\xAA\xE8\x87\x1B\x63\x46"
"\x74\x28\x9D\x05\x30\x20\x62\x41"
"\xC0\x9F\x2C\x36\x2B\x78\xD7\x26"
"\xDF\x58\x51\xED\xFA\xDC\x87\x79"
"\xBF\x8C\xBF\xC4\x0F\xE5\x05\xDA"
"\x45\xE3\x35\x0D\x69\x91\x54\x1C"
"\xE7\x2C\x49\x08\x8B\x72\xFA\x5C"
"\xF1\x6B\xD9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 1011,
.also_non_np = 1,
.np = 2,
.tap = { 1011 - 16, 16 },
}, { /* Generated with Crypto++ */
.key = "\x85\x62\x3F\x1C\xF9\xD6\x1C\xF9"
"\xD6\xB3\x90\x6D\x4A\x90\x6D\x4A"
"\x27\x04\xE1\x27\x04\xE1\xBE\x9B"
"\x78\xBE\x9B\x78\x55\x32\x0F\x55",
.klen = 32,
.iv = "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF"
"\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFD",
.ptext = "\x56\xED\x84\x1B\x8F\x26\xBD\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xC8\x5F\xF6\x6A\x01\x98\x0C\xA3"
"\x3A\xD1\x45\xDC\x73\x0A\x7E\x15"
"\xAC\x20\xB7\x4E\xE5\x59\xF0\x87"
"\x1E\x92\x29\xC0\x34\xCB\x62\xF9"
"\x6D\x04\x9B\x0F\xA6\x3D\xD4\x48"
crypto: testmgr - expand camellia test vectors AVX/AES-NI implementation of camellia cipher processes 16 blocks parallel, so we need to make test vectors larger to check parallel code paths. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-10-20 05:52:46 -06:00
"\xDF\x76\x0D\x81\x18\xAF\x23\xBA"
"\x51\xE8\x5C\xF3\x8A\x21\x95\x2C"
"\xC3\x37\xCE\x65\xFC\x70\x07\x9E"
"\x12\xA9\x40\xD7\x4B\xE2\x79\x10"
"\x84\x1B\xB2\x26\xBD\x54\xEB\x5F"
"\xF6\x8D\x01\x98\x2F\xC6\x3A\xD1"
"\x68\xFF\x73\x0A\xA1\x15\xAC\x43"
"\xDA\x4E\xE5\x7C\x13\x87\x1E\xB5"
"\x29\xC0\x57\xEE\x62\xF9\x90\x04"
"\x9B\x32\xC9\x3D\xD4\x6B\x02\x76"
"\x0D\xA4\x18\xAF\x46\xDD\x51\xE8"
"\x7F\x16\x8A\x21\xB8\x2C\xC3\x5A"
"\xF1\x65\xFC\x93\x07\x9E\x35\xCC"
"\x40\xD7\x6E\x05\x79\x10\xA7\x1B"
"\xB2\x49\xE0\x54\xEB\x82\x19\x8D"
"\x24\xBB\x2F\xC6\x5D\xF4\x68\xFF"
"\x96\x0A\xA1\x38\xCF\x43\xDA\x71"
"\x08\x7C\x13\xAA\x1E\xB5\x4C\xE3"
"\x57\xEE\x85\x1C\x90\x27\xBE\x32"
"\xC9\x60\xF7\x6B\x02\x99\x0D\xA4"
"\x3B\xD2\x46\xDD\x74\x0B\x7F\x16"
"\xAD\x21\xB8\x4F\xE6\x5A\xF1\x88"
"\x1F\x93\x2A\xC1\x35\xCC\x63\xFA"
"\x6E\x05\x9C\x10\xA7\x3E\xD5\x49"
"\xE0\x77\x0E\x82\x19\xB0\x24\xBB"
"\x52\xE9\x5D\xF4\x8B\x22\x96\x2D"
"\xC4\x38\xCF\x66\xFD\x71\x08\x9F"
"\x13\xAA\x41\xD8\x4C\xE3\x7A\x11"
"\x85\x1C\xB3\x27\xBE\x55\xEC\x60"
"\xF7\x8E\x02\x99\x30\xC7\x3B\xD2"
"\x69\x00\x74\x0B\xA2\x16\xAD\x44"
"\xDB\x4F\xE6\x7D\x14\x88\x1F\xB6"
"\x2A\xC1\x58\xEF\x63\xFA\x91\x05"
"\x9C\x33\xCA\x3E\xD5\x6C\x03\x77"
"\x0E\xA5\x19\xB0\x47\xDE\x52\xE9"
"\x80\x17\x8B\x22\xB9\x2D\xC4\x5B"
"\xF2\x66\xFD\x94\x08\x9F\x36\xCD"
"\x41\xD8\x6F\x06\x7A\x11\xA8\x1C"
"\xB3\x4A\xE1\x55\xEC\x83\x1A\x8E"
"\x25\xBC\x30\xC7\x5E\xF5\x69\x00"
"\x97\x0B\xA2\x39\xD0\x44\xDB\x72"
"\x09\x7D\x14\xAB\x1F\xB6\x4D\xE4"
"\x58\xEF\x86\x1D\x91\x28\xBF\x33"
"\xCA\x61\xF8\x6C\x03\x9A\x0E\xA5"
"\x3C\xD3\x47\xDE\x75\x0C\x80\x17"
"\xAE\x22\xB9\x50\xE7\x5B\xF2\x89"
"\x20\x94\x2B\xC2\x36\xCD\x64\xFB"
"\x6F\x06\x9D\x11\xA8\x3F\xD6\x4A"
"\xE1\x78\x0F\x83\x1A\xB1\x25\xBC"
"\x53\xEA\x5E\xF5\x8C\x00\x97\x2E"
"\xC5\x39\xD0\x67\xFE\x72\x09\xA0"
"\x14\xAB\x42\xD9\x4D\xE4\x7B\x12"
"\x86\x1D\xB4\x28\xBF\x56\xED\x61"
"\xF8\x8F\x03\x9A\x31\xC8\x3C\xD3"
"\x6A\x01\x75\x0C\xA3\x17\xAE\x45"
"\xDC\x50\xE7\x7E\x15\x89\x20\xB7"
crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2 Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-13 04:46:35 -06:00
"\x2B\xC2\x59\xF0\x64\xFB\x92\x06"
"\x9D\x34\xCB\x3F\xD6\x6D\x04\x78"
"\x0F\xA6\x1A\xB1\x48\xDF\x53\xEA"
"\x81\x18\x8C\x23\xBA\x2E\xC5\x5C"
"\xF3\x67\xFE\x95\x09\xA0\x37\xCE"
"\x42\xD9\x70\x07\x7B\x12\xA9\x1D"
"\xB4\x4B\xE2\x56\xED\x84\x1B\x8F"
"\x26\xBD\x31\xC8\x5F\xF6\x6A\x01"
"\x98\x0C\xA3\x3A\xD1\x45\xDC\x73"
"\x0A\x7E\x15\xAC\x20\xB7\x4E\xE5"
"\x59\xF0\x87\x1E\x92\x29\xC0\x34"
"\xCB\x62\xF9\x6D\x04\x9B\x0F\xA6"
"\x3D\xD4\x48\xDF\x76\x0D\x81\x18"
"\xAF\x23\xBA\x51\xE8\x5C\xF3\x8A"
"\x21\x95\x2C\xC3\x37\xCE\x65\xFC"
"\x70\x07\x9E\x12\xA9\x40\xD7\x4B"
"\xE2\x79\x10\x84\x1B\xB2\x26\xBD"
"\x54\xEB\x5F\xF6\x8D\x01\x98\x2F"
"\xC6\x3A\xD1\x68\xFF\x73\x0A\xA1"
"\x15\xAC\x43\xDA\x4E\xE5\x7C\x13"
"\x87\x1E\xB5\x29\xC0\x57\xEE\x62"
"\xF9\x90\x04\x9B\x32\xC9\x3D\xD4"
"\x6B\x02\x76\x0D\xA4\x18\xAF\x46"
"\xDD\x51\xE8\x7F\x16\x8A\x21\xB8"
"\x2C\xC3\x5A\xF1\x65\xFC\x93\x07"
"\x9E\x35\xCC\x40\xD7\x6E\x05\x79"
"\x10\xA7\x1B\xB2\x49\xE0\x54\xEB"
"\x82\x19\x8D\x24\xBB\x2F\xC6\x5D"
"\xF4\x68\xFF\x96\x0A\xA1\x38\xCF"
"\x43\xDA\x71\x08\x7C\x13\xAA\x1E"
"\xB5\x4C\xE3\x57\xEE\x85\x1C\x90"
"\x27\xBE\x32\xC9\x60\xF7\x6B\x02"
"\x99\x0D\xA4\x3B\xD2\x46\xDD\x74"
"\x0B\x7F\x16\xAD\x21\xB8\x4F\xE6"
"\x5A\xF1\x88\x1F\x93\x2A\xC1\x35"
"\xCC\x63\xFA\x6E\x05\x9C\x10\xA7"
"\x3E\xD5\x49\xE0\x77\x0E\x82\x19"
"\xB0\x24\xBB\x52\xE9\x5D\xF4\x8B"
"\x22\x96\x2D\xC4\x38\xCF\x66\xFD"
"\x71\x08\x9F\x13\xAA\x41\xD8\x4C"
"\xE3\x7A\x11\x85\x1C\xB3\x27\xBE"
"\x55\xEC\x60\xF7\x8E\x02\x99\x30"
"\xC7\x3B\xD2\x69\x00\x74\x0B\xA2"
"\x16\xAD\x44\xDB\x4F\xE6\x7D\x14"
"\x88\x1F\xB6\x2A\xC1\x58\xEF\x63"
"\xFA\x91\x05\x9C\x33\xCA\x3E\xD5"
"\x6C\x03\x77\x0E\xA5\x19\xB0\x47"
"\xDE\x52\xE9\x80\x17\x8B\x22\xB9"
"\x2D\xC4\x5B\xF2\x66\xFD\x94\x08"
"\x9F\x36\xCD\x41\xD8\x6F\x06\x7A"
"\x11\xA8\x1C\xB3\x4A\xE1\x55\xEC"
"\x83\x1A\x8E\x25\xBC\x30\xC7\x5E"
"\xF5\x69\x00\x97\x0B\xA2\x39\xD0"
"\x44\xDB\x72\x09\x7D\x14\xAB\x1F"
"\xB6\x4D\xE4\x58\xEF\x86\x1D\x91"
"\x28\xBF\x33\xCA\x61\xF8\x6C\x03"
"\x9A\x0E\xA5\x3C\xD3\x47\xDE\x75"
"\x0C\x80\x17\xAE\x22\xB9\x50\xE7"
"\x5B\xF2\x89\x20\x94\x2B\xC2\x36"
"\xCD\x64\xFB\x6F\x06\x9D\x11\xA8"
"\x3F\xD6\x4A\xE1\x78\x0F\x83\x1A"
"\xB1\x25\xBC\x53\xEA\x5E\xF5\x8C"
"\x00\x97\x2E\xC5\x39\xD0\x67\xFE"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\x72\x09\xA0\x14\xAB\x42\xD9\x4D",
.ctext = "\x85\x79\x6C\x8B\x2B\x6D\x14\xF9"
crypto: testmgr - add test vectors for CTR mode IV increasement More precisely, test 'long word' and 'long long word' overflow and carry handling. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:59 -06:00
"\xA6\x83\xB6\x80\x5B\x3A\xF3\x7E"
"\x30\x29\xEB\x1F\xDC\x19\x5F\xEB"
"\xF7\xC4\x27\x04\x51\x87\xD7\x6F"
"\xB8\x4E\x07\xFB\xAC\x3B\x08\xB4"
"\x4D\xCB\xE8\xE1\x71\x7D\x4F\x48"
"\xCD\x81\x64\xA5\xC4\x07\x1A\x9A"
"\x4B\x62\x90\x0E\xC8\xB3\x2B\x6B"
"\x8F\x9C\x6E\x72\x4B\xBA\xEF\x07"
"\x2C\x56\x07\x5E\x37\x30\x60\xA9"
"\xE3\xEF\xD6\x69\xE1\xA1\x77\x64"
"\x93\x75\x7A\xB7\x7A\x3B\xE9\x43"
"\x23\x35\x95\x91\x80\x8A\xC7\xCF"
"\xC3\xD5\xBF\xE7\xFE\x4C\x06\x6B"
"\x05\x19\x48\xE2\x62\xBA\x4F\xF2"
"\xFB\xEE\xE4\xCB\x79\x9D\xA3\x10"
"\x1D\x29\x8C\x1D\x7A\x88\x5A\xDD"
"\x4E\xB6\x18\xAA\xCD\xE6\x33\x96"
"\xD9\x0F\x90\x5A\x78\x76\x4D\x77"
"\x3C\x20\x89\x3B\xA3\xF9\x07\xFD"
"\xE4\xE8\x20\x2D\x15\x0A\x63\x49"
"\xF5\x4F\x89\xD8\xDE\xA1\x28\x78"
"\x28\x07\x09\x1B\x03\x94\x1D\x4B"
"\x82\x28\x1E\x1D\x95\xBA\xAC\x85"
"\x71\x6E\x3C\x18\x4B\x77\x74\x79"
"\xBF\x67\x0A\x53\x3C\x94\xD9\x60"
"\xE9\x6D\x40\x34\xA0\x2A\x53\x5D"
"\x27\xD5\x47\xF9\xC3\x4B\x27\x29"
"\xE4\x76\x9C\x3F\xA7\x1C\x87\xFC"
"\x6E\x0F\xCF\x9B\x60\xF0\xF0\x8B"
"\x70\x1C\x84\x81\x72\x4D\xB4\x98"
"\x23\x62\xE7\x6A\x2B\xFC\xA5\xB2"
"\xFF\xF5\x71\x07\xCD\x90\x23\x13"
"\x19\xD7\x79\x36\x6C\x9D\x55\x8B"
"\x93\x78\x86\x05\x69\x46\xD0\xC5"
"\x39\x09\xEB\x79\xEF\xFA\x9F\xAE"
"\xF3\xD5\x44\xC3\xFD\x86\xD2\x7C"
"\x83\x4B\xD8\x75\x9C\x18\x04\x7B"
"\x73\xAD\x72\xA4\xF6\xAB\xCF\x4B"
"\xCC\x01\x45\x90\xA6\x43\x05\x0C"
"\x6C\x4F\x62\x77\x57\x97\x9F\xEE"
"\x75\xA7\x3C\x38\xD1\x0F\x3D\x0E"
"\x2C\x43\x98\xFB\x13\x65\x73\xE4"
"\x3C\x1E\xD6\x90\x08\xF7\xE0\x99"
"\x3B\xF1\x9D\x6C\x48\xA9\x0E\x32"
"\x17\xC2\xCC\x20\xA1\x19\x26\xAA"
"\xE0\x75\x2F\xFB\x54\x66\x0A\xDF"
"\xB5\xF2\x1F\xC1\x34\x3C\x30\x56"
"\xE8\xDC\xF7\x92\x6B\xBF\x17\x24"
"\xEC\x94\xB5\x3B\xD6\xCE\xA2\x54"
"\x10\x7F\x50\xDE\x69\x77\xD5\x37"
"\xFE\x9C\x10\x83\xC5\xEB\xC9\x53"
"\xB7\xF3\xC4\x20\xAF\x0A\x7E\x57"
"\x3A\xE6\x75\xFE\x89\x00\x6E\x48"
"\xFB\x99\x17\x2C\xF6\x64\x40\x95"
"\x5E\xDC\x7A\xA6\x70\xC7\xF4\xDD"
"\x52\x05\x24\x34\xF9\x0E\xC8\x64"
"\x6D\xE2\xD8\x80\x53\x31\x4C\xFE"
"\xB4\x3A\x5F\x19\xCF\x42\x1B\x22"
"\x0B\x2D\x7B\xF1\xC5\x43\xF7\x5E"
"\x12\xA8\x01\x64\x16\x0B\x26\x5A"
crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2 Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2013-04-13 04:46:35 -06:00
"\x0C\x95\x0F\x40\xC5\x5A\x06\x7C"
"\xCF\xF5\xD5\xB7\x7A\x34\x23\xB6"
"\xAA\x9E\xA8\x98\xA2\xF8\x3D\xD3"
"\x3F\x23\x69\x63\x56\x96\x45\xD6"
"\x74\x23\x1D\x5C\x63\xCC\xD8\x78"
"\x16\xE2\x9C\xD2\x80\x02\xF2\x28"
"\x69\x2F\xC4\xA8\x15\x15\x24\x3B"
"\xCB\xF0\x14\xE4\x62\xC8\xF3\xD1"
"\x03\x58\x1B\x33\x77\x74\x1F\xB4"
"\x07\x86\xF2\x21\xB7\x41\xAE\xBF"
"\x25\xC2\xFF\x51\xEF\xEA\xCE\xC4"
"\x5F\xD9\xB8\x18\x6A\xF0\x0F\x0D"
"\xF8\x04\xBB\x6D\x62\x33\x87\x26"
"\x4F\x2F\x14\x6E\xDC\xDB\x66\x09"
"\x2A\xEF\x7D\x84\x10\xAC\x82\x5E"
"\xD2\xE4\xAD\x74\x7A\x6D\xCC\x3A"
"\x7B\x62\xD8\xD6\x07\x2D\xF7\xDF"
"\x9B\xB3\x82\xCF\x9C\x1D\x76\x5C"
"\xAC\x7B\xD4\x9B\x45\xA1\x64\x11"
"\x66\xF1\xA7\x0B\xF9\xDD\x00\xDD"
"\xA4\x45\x3D\x3E\x03\xC9\x2E\xCB"
"\xC3\x14\x84\x72\xFD\x41\xDC\xBD"
"\x75\xBE\xA8\xE5\x16\x48\x64\x39"
"\xCA\xF3\xE6\xDC\x25\x24\xF1\x6D"
"\xB2\x8D\xC5\x38\x54\xD3\x5D\x6D"
"\x0B\x29\x10\x15\x0E\x13\x3B\xAC"
"\x7E\xCC\x9E\x3E\x18\x48\xA6\x02"
"\xEF\x03\xB2\x2E\xE3\xD2\x70\x21"
"\xB4\x19\x26\xBE\x3A\x3D\x05\xE0"
"\xF8\x09\xAF\xE4\x31\x26\x92\x2F"
"\x8F\x55\xAC\xED\x0B\xB2\xA5\x34"
"\xBE\x50\xB1\x02\x22\x96\xE3\x40"
"\x7B\x70\x50\x6E\x3B\xD5\xE5\xA0"
"\x8E\xA2\xAD\x14\x60\x5C\x7A\x2B"
"\x3D\x1B\x7F\xC1\xC0\x2C\x56\x36"
"\xD2\x0A\x32\x06\x97\x34\xB9\xF4"
"\x6F\x9F\x7E\x80\xD0\x9D\xF7\x6A"
"\x21\xC1\xA2\x6A\xB1\x96\x5B\x4D"
"\x7A\x15\x6C\xC4\x4E\xB8\xE0\x9E"
"\x6C\x50\xF3\x9C\xC9\xB5\x23\xB7"
"\xF1\xD4\x29\x4A\x23\xC4\xAD\x1E"
"\x2C\x07\xD2\x43\x5F\x57\x93\xCA"
"\x85\xF9\x9F\xAD\x4C\xF1\xE4\xB1"
"\x1A\x8E\x28\xA4\xB6\x52\x77\x7E"
"\x68\xC6\x47\xB9\x76\xCC\x65\x5F"
"\x0B\xF9\x67\x93\xD8\x0E\x9A\x37"
"\x5F\x41\xED\x64\x6C\xAD\x5F\xED"
"\x3F\x8D\xFB\x8E\x1E\xA0\xE4\x1F"
"\xC2\xC7\xED\x18\x43\xE1\x20\x86"
"\x5D\xBC\x30\x70\x22\xA1\xDC\x53"
"\x10\x3A\x8D\x47\x82\xCD\x7F\x59"
"\x03\x2D\x6D\xF5\xE7\x79\xD4\x07"
"\x68\x2A\xA5\x42\x19\x4D\xAF\xF5"
"\xED\x47\x83\xBC\x5F\x62\x84\xDA"
"\xDA\x41\xFF\xB0\x1D\x64\xA3\xC8"
"\xBD\x4E\xE0\xB8\x7F\xEE\x55\x0A"
"\x4E\x61\xB2\x51\xF6\x9C\x95\xF6"
"\x92\xBB\xF6\xC5\xF0\x09\x86\xDE"
"\x37\x9E\x29\xF9\x2A\x18\x73\x0D"
"\xDC\x7E\x6B\x7B\x1B\x43\x8C\xEA"
"\x13\xC8\x1A\x47\x0A\x2D\x6D\x56"
"\xCD\xD2\xE7\x53\x1A\xAB\x1C\x3C"
"\xC5\x9B\x03\x70\x29\x2A\x49\x09"
"\x67\xA1\xEA\xD6\x3A\x5B\xBF\x71"
"\x1D\x48\x64\x6C\xFB\xC0\x9E\x36",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 1008,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec camellia_lrw_tv_template[] = {
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
/* Generated from AES-LRW test vectors */
{
.key = "\x45\x62\xac\x25\xf8\x28\x17\x6d"
"\x4c\x26\x84\x14\xb5\x68\x01\x85"
"\x25\x8e\x2a\x05\xe7\x3e\x9d\x03"
"\xee\x5a\x83\x0c\xcc\x09\x4c\x87",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x92\x68\x19\xd7\xb7\x5b\x0a\x31"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x97\xcc\x72\xbe\x99\x17\xeb\x3e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\x59\x70\x47\x14\xf5\x57\x47\x8c"
"\xd7\x79\xe8\x0f\x54\x88\x79\x44"
"\x0d\x48\xf0\xb7\xb1\x5a\x53\xea"
"\x1c\xaa\x6b\x29\xc2\xca\xfb\xaf",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x73\x09\xb7\x50\xb6\x77\x30\x50"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x5c\x8a\x9c\x26\x77\x9d\xfc\x4a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\xd8\x2a\x91\x34\xb2\x6a\x56\x50"
"\x30\xfe\x69\xe2\x37\x7f\x98\x47"
"\xcd\xf9\x0b\x16\x0c\x64\x8f\xb6"
"\xb0\x0d\x0d\x1b\xae\x85\x87\x1f",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x38\x39\x41\x42\x43\x44\x45\x46",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x90\xae\x83\xe0\x22\xb9\x60\x91"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xfa\xa9\xb7\x98\xe3\xed\x87\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\x0f\x6a\xef\xf8\xd3\xd2\xbb\x15"
"\x25\x83\xf7\x3c\x1f\x01\x28\x74"
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
"\xca\xc6\xbc\x35\x4d\x4a\x65\x54"
"\x90\xae\x61\xcf\x7b\xae\xbd\xcc"
"\xad\xe4\x94\xc5\x4a\x29\xae\x70",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x99\xe9\x6e\xd4\xc9\x21\xa5\xf0"
"\xd8\x83\xef\xd9\x07\x16\x5f\x35",
.len = 16,
}, {
.key = "\x8a\xd4\xee\x10\x2f\xbd\x81\xff"
"\xf8\x86\xce\xac\x93\xc5\xad\xc6"
"\xa0\x19\x07\xc0\x9d\xf7\xbb\xdd"
"\x52\x13\xb2\xb7\xf0\xff\x11\xd8"
"\xd6\x08\xd0\xcd\x2e\xb1\x17\x6f",
.klen = 40,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x42\x88\xf4\xcb\x21\x11\x6d\x8e"
"\xde\x1a\xf2\x29\xf1\x4a\xe0\x15",
.len = 16,
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x40\xaa\x34\x86\x4a\x8f\x78\xb9"
"\xdb\xdb\x0f\x3d\x48\x70\xbe\x8d",
.len = 16,
}, {
.key = "\xfb\x76\x15\xb2\x3d\x80\x89\x1d"
"\xd4\x70\x98\x0b\xc7\x95\x84\xc8"
"\xb2\xfb\x64\xce\x60\x97\x87\x8d"
"\x17\xfc\xe4\x5a\x49\xe8\x30\xb7"
"\x6e\x78\x17\xe7\x2d\x5e\x12\xd4"
"\x60\x64\x04\x7a\xf1\x2f\x9e\x0c",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x02\x00\x00\x00\x00",
.ptext = "\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x41\x42\x43\x44\x45\x46",
.ctext = "\x04\xab\x28\x37\x31\x7a\x26\xab"
"\xa1\x70\x1b\x9c\xe7\xdd\x83\xff",
.len = 16,
}, {
.key = "\xf8\xd4\x76\xff\xd6\x46\xee\x6c"
"\x23\x84\xcb\x1c\x77\xd6\x19\x5d"
"\xfe\xf1\xa9\xf3\x7b\xbc\x8d\x21"
"\xa7\x9c\x21\xf8\xcb\x90\x02\x89"
"\xa8\x45\x34\x8e\xc8\xc5\xb5\xf1"
"\x26\xf5\x0e\x76\xfe\xfd\x1b\x1e",
.klen = 48,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.ptext = "\x05\x11\xb7\x18\xab\xc6\x2d\xac"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x70\x5d\xf6\x22\x94\xcd\xe5\x6c"
"\x17\x6b\xf6\x1c\xf0\xf3\x6e\xf8"
"\x50\x38\x1f\x71\x49\xb6\x57\xd6"
"\x8f\xcb\x8d\x6b\xe3\xa6\x29\x90"
"\xfe\x2a\x62\x82\xae\x6d\x8b\xf6"
"\xad\x1e\x9e\x20\x5f\x38\xbe\x04"
"\xda\x10\x8e\xed\xa2\xa4\x87\xab"
"\xda\x6b\xb4\x0c\x75\xba\xd3\x7c"
"\xc9\xac\x42\x31\x95\x7c\xc9\x04"
"\xeb\xd5\x6e\x32\x69\x8a\xdb\xa6"
"\x15\xd7\x3f\x4f\x2f\x66\x69\x03"
"\x9c\x1f\x54\x0f\xde\x1f\xf3\x65"
"\x4c\x96\x12\xed\x7c\x92\x03\x01"
"\x6f\xbc\x35\x93\xac\xf1\x27\xf1"
"\xb4\x96\x82\x5a\x5f\xb0\xa0\x50"
"\x89\xa4\x8e\x66\x44\x85\xcc\xfd"
"\x33\x14\x70\xe3\x96\xb2\xc3\xd3"
"\xbb\x54\x5a\x1a\xf9\x74\xa2\xc5"
"\x2d\x64\x75\xdd\xb4\x54\xe6\x74"
"\x8c\xd3\x9d\x9e\x86\xab\x51\x53"
"\xb7\x93\x3e\x6f\xd0\x4e\x2c\x40"
"\xf6\xa8\x2e\x3e\x9d\xf4\x66\xa5"
"\x76\x12\x73\x44\x1a\x56\xd7\x72"
"\x88\xcd\x21\x8c\x4c\x0f\xfe\xda"
"\x95\xe0\x3a\xa6\xa5\x84\x46\xcd"
"\xd5\x3e\x9d\x3a\xe2\x67\xe6\x60"
"\x1a\xe2\x70\x85\x58\xc2\x1b\x09"
"\xe1\xd7\x2c\xca\xad\xa8\x8f\xf9"
"\xac\xb3\x0e\xdb\xca\x2e\xe2\xb8"
"\x51\x71\xd9\x3c\x6c\xf1\x56\xf8"
"\xea\x9c\xf1\xfb\x0c\xe6\xb7\x10"
"\x1c\xf8\xa9\x7c\xe8\x53\x35\xc1"
"\x90\x3e\x76\x4a\x74\xa4\x21\x2c"
"\xf6\x2c\x4e\x0f\x94\x3a\x88\x2e"
"\x41\x09\x6a\x33\x7d\xf6\xdd\x3f"
"\x8d\x23\x31\x74\x84\xeb\x88\x6e"
"\xcc\xb9\xbc\x22\x83\x19\x07\x22"
"\xa5\x2d\xdf\xa5\xf3\x80\x85\x78"
"\x84\x39\x6a\x6d\x6a\x99\x4f\xa5"
"\x15\xfe\x46\xb0\xe4\x6c\xa5\x41"
"\x3c\xce\x8f\x42\x60\x71\xa7\x75"
"\x08\x40\x65\x8a\x82\xbf\xf5\x43"
"\x71\x96\xa9\x4d\x44\x8a\x20\xbe"
"\xfa\x4d\xbb\xc0\x7d\x31\x96\x65"
"\xe7\x75\xe5\x3e\xfd\x92\x3b\xc9"
"\x55\xbb\x16\x7e\xf7\xc2\x8c\xa4"
"\x40\x1d\xe5\xef\x0e\xdf\xe4\x9a"
"\x62\x73\x65\xfd\x46\x63\x25\x3d"
"\x2b\xaf\xe5\x64\xfe\xa5\x5c\xcf"
"\x24\xf3\xb4\xac\x64\xba\xdf\x4b"
"\xc6\x96\x7d\x81\x2d\x8d\x97\xf7"
"\xc5\x68\x77\x84\x32\x2b\xcc\x85"
"\x74\x96\xf0\x12\x77\x61\xb9\xeb"
"\x71\xaa\x82\xcb\x1c\xdb\x89\xc8"
"\xc6\xb5\xe3\x5c\x7d\x39\x07\x24"
"\xda\x39\x87\x45\xc0\x2b\xbb\x01"
"\xac\xbc\x2a\x5c\x7f\xfc\xe8\xce"
"\x6d\x9c\x6f\xed\xd3\xc1\xa1\xd6"
"\xc5\x55\xa9\x66\x2f\xe1\xc8\x32"
"\xa6\x5d\xa4\x3a\x98\x73\xe8\x45"
"\xa4\xc7\xa8\xb4\xf6\x13\x03\xf6"
"\xe9\x2e\xc4\x29\x0f\x84\xdb\xc4"
"\x21\xc4\xc2\x75\x67\x89\x37\x0a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x90\x69\x8e\xf2\x14\x86\x59\xf9"
"\xec\xe7\xfa\x3f\x48\x9d\x7f\x96"
"\x67\x76\xac\x2c\xd2\x63\x18\x93"
"\x13\xf8\xf1\xf6\x71\x77\xb3\xee"
"\x93\xb2\xcc\xf3\x26\xc1\x16\x4f"
"\xd4\xe8\x43\xc1\x68\xa3\x3e\x06"
"\x38\x51\xff\xa8\xb9\xa4\xeb\xb1"
"\x62\xdd\x78\x81\xea\x1d\xef\x04"
"\x1d\x07\xc1\x67\xc8\xd6\x77\xa1"
"\x84\x95\xf4\x9a\xd9\xbc\x2d\xe2"
"\xf6\x80\xfc\x91\x2a\xbc\x42\xa0"
"\x40\x41\x69\xaa\x71\xc0\x37\xec"
"\x39\xf3\xf2\xec\x82\xc3\x88\x79"
"\xbc\xc3\xaa\xb7\xcf\x6a\x72\x80"
"\x4c\xf4\x84\x8f\x13\x9e\x94\x5c"
"\xe5\xb2\x91\xbb\x92\x51\x4d\xf1"
"\xd6\x0d\x71\x6b\x7a\xc2\x2f\x12"
"\x6f\x75\xc7\x80\x99\x50\x84\xcf"
"\xa8\xeb\xd6\xe1\x1c\x59\x81\x7e"
"\xb9\xb3\xde\x7a\x93\x14\x12\xa2"
"\xf7\x43\xb3\x9d\x1a\x87\x65\x91"
"\x42\x08\x40\x82\x06\x1c\x2d\x55"
"\x6e\x48\xd5\x74\x07\x6e\x9d\x80"
"\xeb\xb4\x97\xa1\x36\xdf\xfa\x74"
"\x79\x7f\x5a\x75\xe7\x71\xc8\x8c"
"\x7e\xf8\x3a\x77\xcd\x32\x05\xf9"
"\x3d\xd4\xe9\xa2\xbb\xc4\x8b\x83"
"\x42\x5c\x82\xfa\xe9\x4b\x96\x3b"
"\x7f\x89\x8b\xf9\xf1\x87\xda\xf0"
"\x87\xef\x13\x5d\xf0\xe2\xc5\xc1"
"\xed\x14\xa9\x57\x19\x63\x40\x04"
"\x24\xeb\x6e\x19\xd1\x3d\x70\x78"
"\xeb\xda\x55\x70\x2c\x4f\x41\x5b"
"\x56\x9f\x1a\xd3\xac\xf1\xc0\xc3"
"\x21\xec\xd7\xd2\x55\x32\x7c\x2e"
"\x3c\x48\x8e\xb4\x85\x35\x47\xfe"
"\xe2\x88\x79\x98\x6a\xc9\x8d\xff"
"\xe9\x89\x6e\xb8\xe2\x97\x00\xbd"
"\xa4\x8f\xba\xd0\x8c\xcb\x79\x99"
"\xb3\xb2\xb2\x7a\xc3\xb7\xef\x75"
"\x23\x52\x76\xc3\x50\x6e\x66\xf8"
"\xa2\xe2\xce\xba\x40\x21\x3f\xc9"
"\x0a\x32\x7f\xf7\x08\x8c\x66\xcf"
"\xd3\xdf\x57\x59\x83\xb8\xe1\x85"
"\xd6\x8f\xfb\x48\x1f\x3a\xc4\x2f"
"\xb4\x2d\x58\xab\xd8\x7f\x5e\x3a"
"\xbc\x62\x3e\xe2\x6a\x52\x0d\x76"
"\x2f\x1c\x1a\x30\xed\x95\x2a\x44"
"\x35\xa5\x83\x04\x84\x01\x99\x56"
"\xb7\xe3\x10\x96\xfa\xdc\x19\xdd"
"\xe2\x7f\xcb\xa0\x49\x1b\xff\x4c"
"\x73\xf6\xbb\x94\x00\xe8\xa9\x3d"
"\xe2\x20\xe9\x3f\xfa\x07\x5d\x77"
"\x06\xd5\x4f\x4d\x02\xb8\x40\x1b"
"\x30\xed\x1a\x50\x19\xef\xc4\x2c"
"\x02\xd9\xc5\xd3\x11\x33\x37\xe5"
"\x2b\xa3\x95\xa6\xee\xd8\x74\x1d"
"\x68\xa0\xeb\xbf\xdd\x5e\x99\x96"
"\x91\xc3\x94\x24\xa5\x12\xa2\x37"
"\xb3\xac\xcf\x2a\xfd\x55\x34\xfe"
"\x79\x92\x3e\xe6\x1b\x49\x57\x5d"
"\x93\x6c\x01\xf7\xcc\x4e\x20\xd1"
"\xb2\x1a\xd8\x4c\xbd\x1d\x10\xe9"
"\x5a\xa8\x92\x7f\xba\xe6\x0c\x95",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec camellia_xts_tv_template[] = {
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
/* Generated from AES-XTS test vectors */
{
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x06\xcb\xa5\xf1\x04\x63\xb2\x41"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\xdc\xca\xfa\x09\xba\x74\xb9\x05"
"\x78\xba\xa4\xf8\x67\x4d\x7e\xad"
"\x20\x18\xf5\x0c\x41\x16\x2a\x61",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\x11\x11\x11\x11\x11\x11\x11\x11"
"\x11\x11\x11\x11\x11\x11\x11\x11"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc2\xb9\xdc\x44\x1d\xdf\xf2\x86"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x8d\x35\x42\x0a\xa5\x5e\x3d\x4f"
"\xb5\x37\x06\xff\xbd\xd4\x91\x70"
"\x80\x1f\xb2\x39\x10\x89\x44\xf5",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\xff\xfe\xfd\xfc\xfb\xfa\xf9\xf8"
"\xf7\xf6\xf5\xf4\xf3\xf2\xf1\xf0"
"\x22\x22\x22\x22\x22\x22\x22\x22"
"\x22\x22\x22\x22\x22\x22\x22\x22",
.klen = 32,
.iv = "\x33\x33\x33\x33\x33\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x44\x44\x44\x44\x44\x44\x44\x44"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44"
"\x44\x44\x44\x44\x44\x44\x44\x44",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x52\x1f\x9d\xf5\x5a\x58\x5a\x7e"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x9f\xd0\x8e\x02\x9c\x9a\x6a\xa7"
"\xb4\x3b\xce\xe7\x17\xaa\x89\x6a"
"\x35\x3c\x6b\xb5\x61\x1c\x79\x38",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xc7\xf9\x0a\xaa\xcb\xb5\x8f\x33"
"\x60\xc3\xe9\x47\x90\xb7\x50\x57"
"\xa3\xad\x81\x2f\xf5\x22\x96\x02"
"\xaa\x7f\xea\xac\x29\x78\xca\x2a"
"\x7c\xcd\x31\x1a\x3c\x40\x0a\x73"
"\x09\x66\xad\x72\x0e\x4d\x5d\x77"
"\xbc\xb8\x76\x80\x37\x59\xa9\x01"
"\x9e\xfb\xdb\x6c\x93\xef\xb6\x8d"
"\x1e\xc1\x94\xa8\xd4\xb5\xb0\x01"
"\xd5\x01\x97\x28\xcd\x7a\x1f\xe8"
"\x08\xda\x76\x00\x65\xcf\x7b\x31"
"\xc6\xfa\xf2\x3b\x00\xa7\x6a\x9e"
"\x6c\x43\x80\x87\xe0\xbb\x4e\xe5"
"\xdc\x8a\xdf\xc3\x1d\x1b\x41\x04"
"\xfb\x54\xdd\x29\x27\xc2\x65\x17"
"\x36\x88\xb0\x85\x8d\x73\x7e\x4b"
"\x1d\x16\x8a\x52\xbc\xa6\xbc\xa4"
"\x8c\xd1\x04\x16\xbf\x8c\x01\x0f"
"\x7e\x6b\x59\x15\x29\xd1\x9b\xd3"
"\x6c\xee\xac\xdc\x45\x58\xca\x5b"
"\x70\x0e\x6a\x12\x86\x82\x79\x9f"
"\x16\xd4\x9d\x67\xcd\x70\x65\x26"
"\x21\x72\x1e\xa1\x94\x8a\x83\x0c"
"\x92\x42\x58\x5e\xa2\xc5\x31\xf3"
"\x7b\xd1\x31\xd4\x15\x80\x31\x61"
"\x5c\x53\x10\xdd\xea\xc8\x83\x5c"
"\x7d\xa7\x05\x66\xcc\x1e\xbb\x05"
"\x47\xae\xb4\x0f\x84\xd8\xf6\xb5"
"\xa1\xc6\x52\x00\x52\xe8\xdc\xd9"
"\x16\x31\xb2\x47\x91\x67\xaa\x28"
"\x2c\x29\x85\xa3\xf7\xf2\x24\x93"
"\x23\x80\x1f\xa8\x1b\x82\x8d\xdc"
"\x9f\x0b\xcd\xb4\x3c\x20\xbc\xec"
"\x4f\xc7\xee\xf8\xfd\xd9\xfb\x7e"
"\x3f\x0d\x23\xfa\x3f\xa7\xcc\x66"
"\x1c\xfe\xa6\x86\xf6\xf7\x85\xc7"
"\x43\xc1\xd4\xfc\xe4\x79\xc9\x1d"
"\xf8\x89\xcd\x20\x27\x84\x5d\x5c"
"\x8e\x4f\x1f\xeb\x08\x21\x4f\xa3"
"\xe0\x7e\x0b\x9c\xe7\x42\xcf\xb7"
"\x3f\x43\xcc\x86\x71\x34\x6a\xd9"
"\x5e\xec\x8f\x36\xc9\x0a\x03\xfe"
"\x18\x41\xdc\x9e\x2e\x75\x20\x3e"
"\xcc\x77\xe0\x8f\xe8\x43\x37\x4c"
"\xed\x1a\x5a\xb3\xfa\x43\xc9\x71"
"\x9f\xc5\xce\xcf\xff\xe7\x77\x1e"
"\x35\x93\xde\x6b\xc0\x6a\x7e\xa9"
"\x34\xb8\x27\x74\x08\xda\xf2\x4a"
"\x23\x5b\x9f\x55\x3a\x57\x82\x52"
"\xea\x6d\xc3\xc7\xf2\xc8\xb5\xdc"
"\xc5\xb9\xbb\xaa\xf2\x29\x9f\x49"
"\x7a\xef\xfe\xdc\x9f\xc9\x28\xe2"
"\x96\x0b\x35\x84\x05\x0d\xd6\x2a"
"\xea\x5a\xbf\x69\xde\xee\x4f\x8f"
"\x84\xb9\xcf\xa7\x57\xea\xe0\xe8"
"\x96\xef\x0f\x0e\xec\xc7\xa6\x74"
"\xb1\xfe\x7a\x6d\x11\xdd\x0e\x15"
"\x4a\x1e\x73\x7f\x55\xea\xf6\xe1"
"\x5b\xb6\x71\xda\xb0\x0c\xba\x26"
"\x5c\x48\x38\x6d\x1c\x32\xb2\x7d"
"\x05\x87\xc2\x1e\x7e\x2d\xd4\x33"
"\xcc\x06\xdb\xe7\x82\x29\x63\xd1"
"\x52\x84\x4f\xee\x27\xe8\x02\xd4"
"\x34\x3c\x69\xc2\xbd\x20\xe6\x7a",
.len = 512,
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
}, {
.key = "\x27\x18\x28\x18\x28\x45\x90\x45"
"\x23\x53\x60\x28\x74\x71\x35\x26"
"\x62\x49\x77\x57\x24\x70\x93\x69"
"\x99\x59\x57\x49\x66\x96\x76\x27"
"\x31\x41\x59\x26\x53\x58\x97\x93"
"\x23\x84\x62\x64\x33\x83\x27\x95"
"\x02\x88\x41\x97\x16\x93\x99\x37"
"\x51\x05\x82\x09\x74\x94\x45\x92",
.klen = 64,
.iv = "\xff\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x49\xcd\xb8\xbf\x2f\x73\x37\x28"
"\x9a\x7f\x6e\x57\x55\xb8\x07\x88"
"\x4a\x0d\x8b\x55\x60\xed\xb6\x7b"
"\xf1\x74\xac\x96\x05\x7b\x32\xca"
"\xd1\x4e\xf1\x58\x29\x16\x24\x6c"
"\xf2\xb3\xe4\x88\x84\xac\x4d\xee"
"\x97\x07\x82\xf0\x07\x12\x38\x0a"
"\x67\x62\xaf\xfd\x85\x9f\x0a\x55"
"\xa5\x20\xc5\x60\xe4\x68\x53\xa4"
"\x0e\x2e\x65\xe3\xe4\x0c\x30\x7c"
"\x1c\x01\x4f\x55\xa9\x13\xeb\x25"
"\x21\x87\xbc\xd3\xe7\x67\x4f\x38"
"\xa8\x14\x25\x71\xe9\x2e\x4c\x21"
"\x41\x82\x0c\x45\x39\x35\xa8\x75"
"\x03\x29\x01\x84\x8c\xab\x48\xbe"
"\x11\x56\x22\x67\xb7\x67\x1a\x09"
"\xa1\x72\x25\x41\x3c\x39\x65\x80"
"\x7d\x2f\xf8\x2c\x73\x04\x58\x9d"
"\xdd\x16\x8b\x63\x70\x4e\xc5\x17"
"\x21\xe0\x84\x51\x4b\x6f\x05\x52"
"\xe3\x63\x34\xfa\xa4\xaf\x33\x20"
"\xc1\xae\x32\xc4\xb8\x2b\xdb\x76"
"\xd9\x02\x31\x2f\xa3\xc6\xd0\x7b"
"\xaf\x1b\x84\xe3\x9b\xbf\xa6\xe0"
"\xb8\x8a\x13\x88\x71\xf4\x11\xa5"
"\xe9\xa9\x10\x33\xe0\xbe\x49\x89"
"\x41\x22\xf5\x9d\x80\x3e\x3b\x76"
"\x01\x16\x50\x6e\x7c\x6a\x81\xe9"
"\x13\x2c\xde\xb2\x5f\x79\xba\xb2"
"\xb1\x75\xae\xd2\x07\x98\x4b\x69"
"\xae\x7d\x5b\x90\xc2\x6c\xe6\x98"
"\xd3\x4c\xa1\xa3\x9c\xc9\x33\x6a"
"\x0d\x23\xb1\x79\x25\x13\x4b\xe5"
"\xaf\x93\x20\x5c\x7f\x06\x7a\x34"
"\x0b\x78\xe3\x67\x26\xe0\xad\x95"
"\xc5\x4e\x26\x22\xcf\x73\x77\x62"
"\x3e\x10\xd7\x90\x4b\x52\x1c\xc9"
"\xef\x38\x52\x18\x0e\x29\x7e\xef"
"\x34\xfe\x31\x95\xc5\xbc\xa8\xe2"
"\xa8\x4e\x9f\xea\xa6\xf0\xfe\x5d"
"\xc5\x39\x86\xed\x2f\x6d\xa0\xfe"
"\x96\xcd\x41\x10\x78\x4e\x0c\xc9"
"\xc3\x6d\x0f\xb7\xe8\xe0\x62\xab"
"\x8b\xf1\x21\x89\xa1\x12\xaa\xfa"
"\x9d\x70\xbe\x4c\xa8\x98\x89\x01"
"\xb9\xe2\x61\xde\x0c\x4a\x0b\xaa"
"\x89\xf5\x14\x79\x18\x8f\x3b\x0d"
"\x21\x17\xf8\x59\x15\x24\x64\x22"
"\x57\x48\x80\xd5\x3d\x92\x30\x07"
"\xd9\xa1\x4a\x23\x16\x43\x48\x0e"
"\x2b\x2d\x1b\x87\xef\x7e\xbd\xfa"
"\x49\xbc\x7e\x68\x6e\xa8\x46\x95"
"\xad\x5e\xfe\x0a\xa8\xd3\x1a\x5d"
"\x6b\x84\xf3\x00\xba\x52\x05\x02"
"\xe3\x96\x4e\xb6\x79\x3f\x43\xd3"
"\x4d\x3f\xd6\xab\x0a\xc4\x75\x2d"
"\xd1\x08\xc3\x6a\xc8\x37\x29\xa0"
"\xcc\x9a\x05\xdd\x5c\xe1\xff\x66"
"\xf2\x7a\x1d\xf2\xaf\xa9\x48\x89"
"\xf5\x21\x0f\x02\x48\x83\x74\xbf"
"\x2e\xe6\x93\x7b\xa0\xf4\xb1\x2b"
"\xb1\x02\x0a\x5c\x79\x19\x3b\x75"
"\xb7\x16\xd8\x12\x5c\xcd\x7d\x4e"
"\xd5\xc6\x99\xcc\x4e\x6c\x94\x95",
.len = 512,
crypto: testmgr - allow non-multi page and multi page skcipher tests from same test template Allow non-multi page and multi page skcipher tests to be run on same test template, to avoid duplicating data. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-09-19 00:42:49 -06:00
.also_non_np = 1,
crypto: testmgr - use chunks smaller than algo block size in chunk tests This patch updates many of the chunked tcrypt test cases so that not all of the chunks are an exact multiple of the block size. This should help uncover cases where the residue passed to blkcipher_walk_done() is incorrect. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-07-27 11:18:50 -06:00
.np = 3,
.tap = { 512 - 20, 4, 16 },
crypto: testmgr - add more camellia test vectors New ECB, CBC, CTR, LRW and XTS test vectors for camellia. Larger ECB/CBC test vectors needed for parallel 2-way camellia implementation. Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-03-05 11:26:21 -07:00
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
/*
* SEED test vectors
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec seed_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = zeroed_string,
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x01\x02\x03\x04\x05\x06\x07"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x5e\xba\xc6\xe0\x05\x4e\x16\x68"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x19\xaf\xf1\xcc\x6d\x34\x6c\xdb",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f",
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = zeroed_string,
.ctext = "\xc1\x1f\x22\xf2\x01\x40\x50\x50"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x84\x48\x35\x97\xe4\x37\x0f\x43",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x47\x06\x48\x08\x51\xe6\x1b\xe8"
"\x5d\x74\xbf\xb3\xfd\x95\x61\x85",
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x83\xa2\xf8\xa2\x88\x64\x1f\xb9"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xa4\xe9\xa5\xcc\x2f\x13\x1c\x7d",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xee\x54\xd1\x3e\xbc\xae\x70\x6d"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x22\x6b\xc3\x14\x2c\xd4\x0d\x4a",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.key = "\x28\xdb\xc3\xbc\x49\xff\xd8\x7d"
"\xcf\xa5\x09\xb1\x1d\x42\x2b\xe7",
.klen = 16,
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\xb4\x1e\x6b\xe2\xeb\xa8\x4a\x14"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x8e\x2e\xed\x84\x59\x3c\x5e\xc7",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x9b\x9b\x7b\xfc\xd1\x81\x3c\xb9"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x5d\x0b\x36\x18\xf4\x0f\x51\x22",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 16,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec salsa20_stream_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* Testvectors from verified.test-vectors submitted to ECRYPT.
* They are truncated to size 39, 64, 111, 129 to test a variety
* of input length.
*/
{ /* Set 3, vector 0 */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0A\x0B\x0C\x0D\x0E\x0F",
.klen = 16,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x2D\xD5\xC3\xF7\xBA\x2B\x20\xF7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x68\x02\x41\x0C\x68\x86\x88\x89"
"\x5A\xD8\xC1\xBD\x4E\xA6\xC9\xB1"
"\x40\xFB\x9B\x90\xE2\x10\x49\xBF"
"\x58\x3F\x52\x79\x70\xEB\xC1",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 39,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Set 5, vector 0 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 16,
.iv = "\x80\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xB6\x6C\x1E\x44\x46\xDD\x95\x57"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xE5\x78\xE2\x23\xB0\xB7\x68\x01"
"\x7B\x23\xB2\x67\xBB\x02\x34\xAE"
"\x46\x26\xBF\x44\x3F\x21\x97\x76"
"\x43\x6F\xB1\x9F\xD0\xE8\x86\x6F"
"\xCD\x0D\xE9\xA9\x53\x8F\x4A\x09"
"\xCA\x9A\xC0\x73\x2E\x30\xBC\xF9"
"\x8E\x4F\x13\xE4\xB9\xE2\x01\xD9",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 64,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Set 3, vector 27 */
.key = "\x1B\x1C\x1D\x1E\x1F\x20\x21\x22"
"\x23\x24\x25\x26\x27\x28\x29\x2A"
"\x2B\x2C\x2D\x2E\x2F\x30\x31\x32"
"\x33\x34\x35\x36\x37\x38\x39\x3A",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xAE\x39\x50\x8E\xAC\x9A\xEC\xE7"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xBF\x97\xBB\x20\xB9\xDE\xE4\x1F"
"\x87\xD9\x47\xF8\x28\x91\x35\x98"
"\xDB\x72\xCC\x23\x29\x48\x56\x5E"
"\x83\x7E\x0B\xF3\x7D\x5D\x38\x7B"
"\x2D\x71\x02\xB4\x3B\xB5\xD8\x23"
"\xB0\x4A\xDF\x3C\xEC\xB6\xD9\x3B"
"\x9B\xA7\x52\xBE\xC5\xD4\x50\x59"
"\x15\x14\xB4\x0E\x40\xE6\x53\xD1"
"\x83\x9C\x5B\xA0\x92\x29\x6B\x5E"
"\x96\x5B\x1E\x2F\xD3\xAC\xC1\x92"
"\xB1\x41\x3F\x19\x2F\xC4\x3B\xC6"
"\x95\x46\x45\x54\xE9\x75\x03\x08"
"\x44\xAF\xE5\x8A\x81\x12\x09",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 111,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* Set 5, vector 27 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.iv = "\x00\x00\x00\x10\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xD2\xDB\x1A\x5C\xF1\xC1\xAC\xDB"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xE8\x1A\x7A\x43\x40\xEF\x53\x43"
"\x5E\x7F\x4B\x1A\x50\x52\x3F\x8D"
"\x28\x3D\xCF\x85\x1D\x69\x6E\x60"
"\xF2\xDE\x74\x56\x18\x1B\x84\x10"
"\xD4\x62\xBA\x60\x50\xF0\x61\xF2"
"\x1C\x78\x7F\xC1\x24\x34\xAF\x58"
"\xBF\x2C\x59\xCA\x90\x77\xF3\xB0"
"\x5B\x4A\xDF\x89\xCE\x2C\x2F\xFC"
"\x67\xF0\xE3\x45\xE8\xB3\xB3\x75"
"\xA0\x95\x71\xA1\x29\x39\x94\xCA"
"\x45\x2F\xBD\xCB\x10\xB6\xBE\x9F"
"\x8E\xF9\xB2\x01\x0A\x5A\x0A\xB7"
"\x6B\x9D\x70\x8E\x4B\xD6\x2F\xCD"
"\x2E\x40\x48\x75\xE9\xE2\x21\x45"
"\x0B\xC9\xB6\xB5\x66\xBC\x9A\x59"
"\x5A",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 129,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, { /* large test vector generated using Crypto++ */
.key = "\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext =
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x00\x01\x02\x03\x04\x05\x06\x07"
"\x08\x09\x0a\x0b\x0c\x0d\x0e\x0f"
"\x10\x11\x12\x13\x14\x15\x16\x17"
"\x18\x19\x1a\x1b\x1c\x1d\x1e\x1f"
"\x20\x21\x22\x23\x24\x25\x26\x27"
"\x28\x29\x2a\x2b\x2c\x2d\x2e\x2f"
"\x30\x31\x32\x33\x34\x35\x36\x37"
"\x38\x39\x3a\x3b\x3c\x3d\x3e\x3f"
"\x40\x41\x42\x43\x44\x45\x46\x47"
"\x48\x49\x4a\x4b\x4c\x4d\x4e\x4f"
"\x50\x51\x52\x53\x54\x55\x56\x57"
"\x58\x59\x5a\x5b\x5c\x5d\x5e\x5f"
"\x60\x61\x62\x63\x64\x65\x66\x67"
"\x68\x69\x6a\x6b\x6c\x6d\x6e\x6f"
"\x70\x71\x72\x73\x74\x75\x76\x77"
"\x78\x79\x7a\x7b\x7c\x7d\x7e\x7f"
"\x80\x81\x82\x83\x84\x85\x86\x87"
"\x88\x89\x8a\x8b\x8c\x8d\x8e\x8f"
"\x90\x91\x92\x93\x94\x95\x96\x97"
"\x98\x99\x9a\x9b\x9c\x9d\x9e\x9f"
"\xa0\xa1\xa2\xa3\xa4\xa5\xa6\xa7"
"\xa8\xa9\xaa\xab\xac\xad\xae\xaf"
"\xb0\xb1\xb2\xb3\xb4\xb5\xb6\xb7"
"\xb8\xb9\xba\xbb\xbc\xbd\xbe\xbf"
"\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7"
"\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf"
"\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd7"
"\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf"
"\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7"
"\xe8\xe9\xea\xeb\xec\xed\xee\xef"
"\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf7"
"\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff"
"\x00\x03\x06\x09\x0c\x0f\x12\x15"
"\x18\x1b\x1e\x21\x24\x27\x2a\x2d"
"\x30\x33\x36\x39\x3c\x3f\x42\x45"
"\x48\x4b\x4e\x51\x54\x57\x5a\x5d"
"\x60\x63\x66\x69\x6c\x6f\x72\x75"
"\x78\x7b\x7e\x81\x84\x87\x8a\x8d"
"\x90\x93\x96\x99\x9c\x9f\xa2\xa5"
"\xa8\xab\xae\xb1\xb4\xb7\xba\xbd"
"\xc0\xc3\xc6\xc9\xcc\xcf\xd2\xd5"
"\xd8\xdb\xde\xe1\xe4\xe7\xea\xed"
"\xf0\xf3\xf6\xf9\xfc\xff\x02\x05"
"\x08\x0b\x0e\x11\x14\x17\x1a\x1d"
"\x20\x23\x26\x29\x2c\x2f\x32\x35"
"\x38\x3b\x3e\x41\x44\x47\x4a\x4d"
"\x50\x53\x56\x59\x5c\x5f\x62\x65"
"\x68\x6b\x6e\x71\x74\x77\x7a\x7d"
"\x80\x83\x86\x89\x8c\x8f\x92\x95"
"\x98\x9b\x9e\xa1\xa4\xa7\xaa\xad"
"\xb0\xb3\xb6\xb9\xbc\xbf\xc2\xc5"
"\xc8\xcb\xce\xd1\xd4\xd7\xda\xdd"
"\xe0\xe3\xe6\xe9\xec\xef\xf2\xf5"
"\xf8\xfb\xfe\x01\x04\x07\x0a\x0d"
"\x10\x13\x16\x19\x1c\x1f\x22\x25"
"\x28\x2b\x2e\x31\x34\x37\x3a\x3d"
"\x40\x43\x46\x49\x4c\x4f\x52\x55"
"\x58\x5b\x5e\x61\x64\x67\x6a\x6d"
"\x70\x73\x76\x79\x7c\x7f\x82\x85"
"\x88\x8b\x8e\x91\x94\x97\x9a\x9d"
"\xa0\xa3\xa6\xa9\xac\xaf\xb2\xb5"
"\xb8\xbb\xbe\xc1\xc4\xc7\xca\xcd"
"\xd0\xd3\xd6\xd9\xdc\xdf\xe2\xe5"
"\xe8\xeb\xee\xf1\xf4\xf7\xfa\xfd"
"\x00\x05\x0a\x0f\x14\x19\x1e\x23"
"\x28\x2d\x32\x37\x3c\x41\x46\x4b"
"\x50\x55\x5a\x5f\x64\x69\x6e\x73"
"\x78\x7d\x82\x87\x8c\x91\x96\x9b"
"\xa0\xa5\xaa\xaf\xb4\xb9\xbe\xc3"
"\xc8\xcd\xd2\xd7\xdc\xe1\xe6\xeb"
"\xf0\xf5\xfa\xff\x04\x09\x0e\x13"
"\x18\x1d\x22\x27\x2c\x31\x36\x3b"
"\x40\x45\x4a\x4f\x54\x59\x5e\x63"
"\x68\x6d\x72\x77\x7c\x81\x86\x8b"
"\x90\x95\x9a\x9f\xa4\xa9\xae\xb3"
"\xb8\xbd\xc2\xc7\xcc\xd1\xd6\xdb"
"\xe0\xe5\xea\xef\xf4\xf9\xfe\x03"
"\x08\x0d\x12\x17\x1c\x21\x26\x2b"
"\x30\x35\x3a\x3f\x44\x49\x4e\x53"
"\x58\x5d\x62\x67\x6c\x71\x76\x7b"
"\x80\x85\x8a\x8f\x94\x99\x9e\xa3"
"\xa8\xad\xb2\xb7\xbc\xc1\xc6\xcb"
"\xd0\xd5\xda\xdf\xe4\xe9\xee\xf3"
"\xf8\xfd\x02\x07\x0c\x11\x16\x1b"
"\x20\x25\x2a\x2f\x34\x39\x3e\x43"
"\x48\x4d\x52\x57\x5c\x61\x66\x6b"
"\x70\x75\x7a\x7f\x84\x89\x8e\x93"
"\x98\x9d\xa2\xa7\xac\xb1\xb6\xbb"
"\xc0\xc5\xca\xcf\xd4\xd9\xde\xe3"
"\xe8\xed\xf2\xf7\xfc\x01\x06\x0b"
"\x10\x15\x1a\x1f\x24\x29\x2e\x33"
"\x38\x3d\x42\x47\x4c\x51\x56\x5b"
"\x60\x65\x6a\x6f\x74\x79\x7e\x83"
"\x88\x8d\x92\x97\x9c\xa1\xa6\xab"
"\xb0\xb5\xba\xbf\xc4\xc9\xce\xd3"
"\xd8\xdd\xe2\xe7\xec\xf1\xf6\xfb"
"\x00\x07\x0e\x15\x1c\x23\x2a\x31"
"\x38\x3f\x46\x4d\x54\x5b\x62\x69"
"\x70\x77\x7e\x85\x8c\x93\x9a\xa1"
"\xa8\xaf\xb6\xbd\xc4\xcb\xd2\xd9"
"\xe0\xe7\xee\xf5\xfc\x03\x0a\x11"
"\x18\x1f\x26\x2d\x34\x3b\x42\x49"
"\x50\x57\x5e\x65\x6c\x73\x7a\x81"
"\x88\x8f\x96\x9d\xa4\xab\xb2\xb9"
"\xc0\xc7\xce\xd5\xdc\xe3\xea\xf1"
"\xf8\xff\x06\x0d\x14\x1b\x22\x29"
"\x30\x37\x3e\x45\x4c\x53\x5a\x61"
"\x68\x6f\x76\x7d\x84\x8b\x92\x99"
"\xa0\xa7\xae\xb5\xbc\xc3\xca\xd1"
"\xd8\xdf\xe6\xed\xf4\xfb\x02\x09"
"\x10\x17\x1e\x25\x2c\x33\x3a\x41"
"\x48\x4f\x56\x5d\x64\x6b\x72\x79"
"\x80\x87\x8e\x95\x9c\xa3\xaa\xb1"
"\xb8\xbf\xc6\xcd\xd4\xdb\xe2\xe9"
"\xf0\xf7\xfe\x05\x0c\x13\x1a\x21"
"\x28\x2f\x36\x3d\x44\x4b\x52\x59"
"\x60\x67\x6e\x75\x7c\x83\x8a\x91"
"\x98\x9f\xa6\xad\xb4\xbb\xc2\xc9"
"\xd0\xd7\xde\xe5\xec\xf3\xfa\x01"
"\x08\x0f\x16\x1d\x24\x2b\x32\x39"
"\x40\x47\x4e\x55\x5c\x63\x6a\x71"
"\x78\x7f\x86\x8d\x94\x9b\xa2\xa9"
"\xb0\xb7\xbe\xc5\xcc\xd3\xda\xe1"
"\xe8\xef\xf6\xfd\x04\x0b\x12\x19"
"\x20\x27\x2e\x35\x3c\x43\x4a\x51"
"\x58\x5f\x66\x6d\x74\x7b\x82\x89"
"\x90\x97\x9e\xa5\xac\xb3\xba\xc1"
"\xc8\xcf\xd6\xdd\xe4\xeb\xf2\xf9"
"\x00\x09\x12\x1b\x24\x2d\x36\x3f"
"\x48\x51\x5a\x63\x6c\x75\x7e\x87"
"\x90\x99\xa2\xab\xb4\xbd\xc6\xcf"
"\xd8\xe1\xea\xf3\xfc\x05\x0e\x17"
"\x20\x29\x32\x3b\x44\x4d\x56\x5f"
"\x68\x71\x7a\x83\x8c\x95\x9e\xa7"
"\xb0\xb9\xc2\xcb\xd4\xdd\xe6\xef"
"\xf8\x01\x0a\x13\x1c\x25\x2e\x37"
"\x40\x49\x52\x5b\x64\x6d\x76\x7f"
"\x88\x91\x9a\xa3\xac\xb5\xbe\xc7"
"\xd0\xd9\xe2\xeb\xf4\xfd\x06\x0f"
"\x18\x21\x2a\x33\x3c\x45\x4e\x57"
"\x60\x69\x72\x7b\x84\x8d\x96\x9f"
"\xa8\xb1\xba\xc3\xcc\xd5\xde\xe7"
"\xf0\xf9\x02\x0b\x14\x1d\x26\x2f"
"\x38\x41\x4a\x53\x5c\x65\x6e\x77"
"\x80\x89\x92\x9b\xa4\xad\xb6\xbf"
"\xc8\xd1\xda\xe3\xec\xf5\xfe\x07"
"\x10\x19\x22\x2b\x34\x3d\x46\x4f"
"\x58\x61\x6a\x73\x7c\x85\x8e\x97"
"\xa0\xa9\xb2\xbb\xc4\xcd\xd6\xdf"
"\xe8\xf1\xfa\x03\x0c\x15\x1e\x27"
"\x30\x39\x42\x4b\x54\x5d\x66\x6f"
"\x78\x81\x8a\x93\x9c\xa5\xae\xb7"
"\xc0\xc9\xd2\xdb\xe4\xed\xf6\xff"
"\x08\x11\x1a\x23\x2c\x35\x3e\x47"
"\x50\x59\x62\x6b\x74\x7d\x86\x8f"
"\x98\xa1\xaa\xb3\xbc\xc5\xce\xd7"
"\xe0\xe9\xf2\xfb\x04\x0d\x16\x1f"
"\x28\x31\x3a\x43\x4c\x55\x5e\x67"
"\x70\x79\x82\x8b\x94\x9d\xa6\xaf"
"\xb8\xc1\xca\xd3\xdc\xe5\xee\xf7"
"\x00\x0b\x16\x21\x2c\x37\x42\x4d"
"\x58\x63\x6e\x79\x84\x8f\x9a\xa5"
"\xb0\xbb\xc6\xd1\xdc\xe7\xf2\xfd"
"\x08\x13\x1e\x29\x34\x3f\x4a\x55"
"\x60\x6b\x76\x81\x8c\x97\xa2\xad"
"\xb8\xc3\xce\xd9\xe4\xef\xfa\x05"
"\x10\x1b\x26\x31\x3c\x47\x52\x5d"
"\x68\x73\x7e\x89\x94\x9f\xaa\xb5"
"\xc0\xcb\xd6\xe1\xec\xf7\x02\x0d"
"\x18\x23\x2e\x39\x44\x4f\x5a\x65"
"\x70\x7b\x86\x91\x9c\xa7\xb2\xbd"
"\xc8\xd3\xde\xe9\xf4\xff\x0a\x15"
"\x20\x2b\x36\x41\x4c\x57\x62\x6d"
"\x78\x83\x8e\x99\xa4\xaf\xba\xc5"
"\xd0\xdb\xe6\xf1\xfc\x07\x12\x1d"
"\x28\x33\x3e\x49\x54\x5f\x6a\x75"
"\x80\x8b\x96\xa1\xac\xb7\xc2\xcd"
"\xd8\xe3\xee\xf9\x04\x0f\x1a\x25"
"\x30\x3b\x46\x51\x5c\x67\x72\x7d"
"\x88\x93\x9e\xa9\xb4\xbf\xca\xd5"
"\xe0\xeb\xf6\x01\x0c\x17\x22\x2d"
"\x38\x43\x4e\x59\x64\x6f\x7a\x85"
"\x90\x9b\xa6\xb1\xbc\xc7\xd2\xdd"
"\xe8\xf3\xfe\x09\x14\x1f\x2a\x35"
"\x40\x4b\x56\x61\x6c\x77\x82\x8d"
"\x98\xa3\xae\xb9\xc4\xcf\xda\xe5"
"\xf0\xfb\x06\x11\x1c\x27\x32\x3d"
"\x48\x53\x5e\x69\x74\x7f\x8a\x95"
"\xa0\xab\xb6\xc1\xcc\xd7\xe2\xed"
"\xf8\x03\x0e\x19\x24\x2f\x3a\x45"
"\x50\x5b\x66\x71\x7c\x87\x92\x9d"
"\xa8\xb3\xbe\xc9\xd4\xdf\xea\xf5"
"\x00\x0d\x1a\x27\x34\x41\x4e\x5b"
"\x68\x75\x82\x8f\x9c\xa9\xb6\xc3"
"\xd0\xdd\xea\xf7\x04\x11\x1e\x2b"
"\x38\x45\x52\x5f\x6c\x79\x86\x93"
"\xa0\xad\xba\xc7\xd4\xe1\xee\xfb"
"\x08\x15\x22\x2f\x3c\x49\x56\x63"
"\x70\x7d\x8a\x97\xa4\xb1\xbe\xcb"
"\xd8\xe5\xf2\xff\x0c\x19\x26\x33"
"\x40\x4d\x5a\x67\x74\x81\x8e\x9b"
"\xa8\xb5\xc2\xcf\xdc\xe9\xf6\x03"
"\x10\x1d\x2a\x37\x44\x51\x5e\x6b"
"\x78\x85\x92\x9f\xac\xb9\xc6\xd3"
"\xe0\xed\xfa\x07\x14\x21\x2e\x3b"
"\x48\x55\x62\x6f\x7c\x89\x96\xa3"
"\xb0\xbd\xca\xd7\xe4\xf1\xfe\x0b"
"\x18\x25\x32\x3f\x4c\x59\x66\x73"
"\x80\x8d\x9a\xa7\xb4\xc1\xce\xdb"
"\xe8\xf5\x02\x0f\x1c\x29\x36\x43"
"\x50\x5d\x6a\x77\x84\x91\x9e\xab"
"\xb8\xc5\xd2\xdf\xec\xf9\x06\x13"
"\x20\x2d\x3a\x47\x54\x61\x6e\x7b"
"\x88\x95\xa2\xaf\xbc\xc9\xd6\xe3"
"\xf0\xfd\x0a\x17\x24\x31\x3e\x4b"
"\x58\x65\x72\x7f\x8c\x99\xa6\xb3"
"\xc0\xcd\xda\xe7\xf4\x01\x0e\x1b"
"\x28\x35\x42\x4f\x5c\x69\x76\x83"
"\x90\x9d\xaa\xb7\xc4\xd1\xde\xeb"
"\xf8\x05\x12\x1f\x2c\x39\x46\x53"
"\x60\x6d\x7a\x87\x94\xa1\xae\xbb"
"\xc8\xd5\xe2\xef\xfc\x09\x16\x23"
"\x30\x3d\x4a\x57\x64\x71\x7e\x8b"
"\x98\xa5\xb2\xbf\xcc\xd9\xe6\xf3"
"\x00\x0f\x1e\x2d\x3c\x4b\x5a\x69"
"\x78\x87\x96\xa5\xb4\xc3\xd2\xe1"
"\xf0\xff\x0e\x1d\x2c\x3b\x4a\x59"
"\x68\x77\x86\x95\xa4\xb3\xc2\xd1"
"\xe0\xef\xfe\x0d\x1c\x2b\x3a\x49"
"\x58\x67\x76\x85\x94\xa3\xb2\xc1"
"\xd0\xdf\xee\xfd\x0c\x1b\x2a\x39"
"\x48\x57\x66\x75\x84\x93\xa2\xb1"
"\xc0\xcf\xde\xed\xfc\x0b\x1a\x29"
"\x38\x47\x56\x65\x74\x83\x92\xa1"
"\xb0\xbf\xce\xdd\xec\xfb\x0a\x19"
"\x28\x37\x46\x55\x64\x73\x82\x91"
"\xa0\xaf\xbe\xcd\xdc\xeb\xfa\x09"
"\x18\x27\x36\x45\x54\x63\x72\x81"
"\x90\x9f\xae\xbd\xcc\xdb\xea\xf9"
"\x08\x17\x26\x35\x44\x53\x62\x71"
"\x80\x8f\x9e\xad\xbc\xcb\xda\xe9"
"\xf8\x07\x16\x25\x34\x43\x52\x61"
"\x70\x7f\x8e\x9d\xac\xbb\xca\xd9"
"\xe8\xf7\x06\x15\x24\x33\x42\x51"
"\x60\x6f\x7e\x8d\x9c\xab\xba\xc9"
"\xd8\xe7\xf6\x05\x14\x23\x32\x41"
"\x50\x5f\x6e\x7d\x8c\x9b\xaa\xb9"
"\xc8\xd7\xe6\xf5\x04\x13\x22\x31"
"\x40\x4f\x5e\x6d\x7c\x8b\x9a\xa9"
"\xb8\xc7\xd6\xe5\xf4\x03\x12\x21"
"\x30\x3f\x4e\x5d\x6c\x7b\x8a\x99"
"\xa8\xb7\xc6\xd5\xe4\xf3\x02\x11"
"\x20\x2f\x3e\x4d\x5c\x6b\x7a\x89"
"\x98\xa7\xb6\xc5\xd4\xe3\xf2\x01"
"\x10\x1f\x2e\x3d\x4c\x5b\x6a\x79"
"\x88\x97\xa6\xb5\xc4\xd3\xe2\xf1"
"\x00\x11\x22\x33\x44\x55\x66\x77"
"\x88\x99\xaa\xbb\xcc\xdd\xee\xff"
"\x10\x21\x32\x43\x54\x65\x76\x87"
"\x98\xa9\xba\xcb\xdc\xed\xfe\x0f"
"\x20\x31\x42\x53\x64\x75\x86\x97"
"\xa8\xb9\xca\xdb\xec\xfd\x0e\x1f"
"\x30\x41\x52\x63\x74\x85\x96\xa7"
"\xb8\xc9\xda\xeb\xfc\x0d\x1e\x2f"
"\x40\x51\x62\x73\x84\x95\xa6\xb7"
"\xc8\xd9\xea\xfb\x0c\x1d\x2e\x3f"
"\x50\x61\x72\x83\x94\xa5\xb6\xc7"
"\xd8\xe9\xfa\x0b\x1c\x2d\x3e\x4f"
"\x60\x71\x82\x93\xa4\xb5\xc6\xd7"
"\xe8\xf9\x0a\x1b\x2c\x3d\x4e\x5f"
"\x70\x81\x92\xa3\xb4\xc5\xd6\xe7"
"\xf8\x09\x1a\x2b\x3c\x4d\x5e\x6f"
"\x80\x91\xa2\xb3\xc4\xd5\xe6\xf7"
"\x08\x19\x2a\x3b\x4c\x5d\x6e\x7f"
"\x90\xa1\xb2\xc3\xd4\xe5\xf6\x07"
"\x18\x29\x3a\x4b\x5c\x6d\x7e\x8f"
"\xa0\xb1\xc2\xd3\xe4\xf5\x06\x17"
"\x28\x39\x4a\x5b\x6c\x7d\x8e\x9f"
"\xb0\xc1\xd2\xe3\xf4\x05\x16\x27"
"\x38\x49\x5a\x6b\x7c\x8d\x9e\xaf"
"\xc0\xd1\xe2\xf3\x04\x15\x26\x37"
"\x48\x59\x6a\x7b\x8c\x9d\xae\xbf"
"\xd0\xe1\xf2\x03\x14\x25\x36\x47"
"\x58\x69\x7a\x8b\x9c\xad\xbe\xcf"
"\xe0\xf1\x02\x13\x24\x35\x46\x57"
"\x68\x79\x8a\x9b\xac\xbd\xce\xdf"
"\xf0\x01\x12\x23\x34\x45\x56\x67"
"\x78\x89\x9a\xab\xbc\xcd\xde\xef"
"\x00\x13\x26\x39\x4c\x5f\x72\x85"
"\x98\xab\xbe\xd1\xe4\xf7\x0a\x1d"
"\x30\x43\x56\x69\x7c\x8f\xa2\xb5"
"\xc8\xdb\xee\x01\x14\x27\x3a\x4d"
"\x60\x73\x86\x99\xac\xbf\xd2\xe5"
"\xf8\x0b\x1e\x31\x44\x57\x6a\x7d"
"\x90\xa3\xb6\xc9\xdc\xef\x02\x15"
"\x28\x3b\x4e\x61\x74\x87\x9a\xad"
"\xc0\xd3\xe6\xf9\x0c\x1f\x32\x45"
"\x58\x6b\x7e\x91\xa4\xb7\xca\xdd"
"\xf0\x03\x16\x29\x3c\x4f\x62\x75"
"\x88\x9b\xae\xc1\xd4\xe7\xfa\x0d"
"\x20\x33\x46\x59\x6c\x7f\x92\xa5"
"\xb8\xcb\xde\xf1\x04\x17\x2a\x3d"
"\x50\x63\x76\x89\x9c\xaf\xc2\xd5"
"\xe8\xfb\x0e\x21\x34\x47\x5a\x6d"
"\x80\x93\xa6\xb9\xcc\xdf\xf2\x05"
"\x18\x2b\x3e\x51\x64\x77\x8a\x9d"
"\xb0\xc3\xd6\xe9\xfc\x0f\x22\x35"
"\x48\x5b\x6e\x81\x94\xa7\xba\xcd"
"\xe0\xf3\x06\x19\x2c\x3f\x52\x65"
"\x78\x8b\x9e\xb1\xc4\xd7\xea\xfd"
"\x10\x23\x36\x49\x5c\x6f\x82\x95"
"\xa8\xbb\xce\xe1\xf4\x07\x1a\x2d"
"\x40\x53\x66\x79\x8c\x9f\xb2\xc5"
"\xd8\xeb\xfe\x11\x24\x37\x4a\x5d"
"\x70\x83\x96\xa9\xbc\xcf\xe2\xf5"
"\x08\x1b\x2e\x41\x54\x67\x7a\x8d"
"\xa0\xb3\xc6\xd9\xec\xff\x12\x25"
"\x38\x4b\x5e\x71\x84\x97\xaa\xbd"
"\xd0\xe3\xf6\x09\x1c\x2f\x42\x55"
"\x68\x7b\x8e\xa1\xb4\xc7\xda\xed"
"\x00\x15\x2a\x3f\x54\x69\x7e\x93"
"\xa8\xbd\xd2\xe7\xfc\x11\x26\x3b"
"\x50\x65\x7a\x8f\xa4\xb9\xce\xe3"
"\xf8\x0d\x22\x37\x4c\x61\x76\x8b"
"\xa0\xb5\xca\xdf\xf4\x09\x1e\x33"
"\x48\x5d\x72\x87\x9c\xb1\xc6\xdb"
"\xf0\x05\x1a\x2f\x44\x59\x6e\x83"
"\x98\xad\xc2\xd7\xec\x01\x16\x2b"
"\x40\x55\x6a\x7f\x94\xa9\xbe\xd3"
"\xe8\xfd\x12\x27\x3c\x51\x66\x7b"
"\x90\xa5\xba\xcf\xe4\xf9\x0e\x23"
"\x38\x4d\x62\x77\x8c\xa1\xb6\xcb"
"\xe0\xf5\x0a\x1f\x34\x49\x5e\x73"
"\x88\x9d\xb2\xc7\xdc\xf1\x06\x1b"
"\x30\x45\x5a\x6f\x84\x99\xae\xc3"
"\xd8\xed\x02\x17\x2c\x41\x56\x6b"
"\x80\x95\xaa\xbf\xd4\xe9\xfe\x13"
"\x28\x3d\x52\x67\x7c\x91\xa6\xbb"
"\xd0\xe5\xfa\x0f\x24\x39\x4e\x63"
"\x78\x8d\xa2\xb7\xcc\xe1\xf6\x0b"
"\x20\x35\x4a\x5f\x74\x89\x9e\xb3"
"\xc8\xdd\xf2\x07\x1c\x31\x46\x5b"
"\x70\x85\x9a\xaf\xc4\xd9\xee\x03"
"\x18\x2d\x42\x57\x6c\x81\x96\xab"
"\xc0\xd5\xea\xff\x14\x29\x3e\x53"
"\x68\x7d\x92\xa7\xbc\xd1\xe6\xfb"
"\x10\x25\x3a\x4f\x64\x79\x8e\xa3"
"\xb8\xcd\xe2\xf7\x0c\x21\x36\x4b"
"\x60\x75\x8a\x9f\xb4\xc9\xde\xf3"
"\x08\x1d\x32\x47\x5c\x71\x86\x9b"
"\xb0\xc5\xda\xef\x04\x19\x2e\x43"
"\x58\x6d\x82\x97\xac\xc1\xd6\xeb"
"\x00\x17\x2e\x45\x5c\x73\x8a\xa1"
"\xb8\xcf\xe6\xfd\x14\x2b\x42\x59"
"\x70\x87\x9e\xb5\xcc\xe3\xfa\x11"
"\x28\x3f\x56\x6d\x84\x9b\xb2\xc9"
"\xe0\xf7\x0e\x25\x3c\x53\x6a\x81"
"\x98\xaf\xc6\xdd\xf4\x0b\x22\x39"
"\x50\x67\x7e\x95\xac\xc3\xda\xf1"
"\x08\x1f\x36\x4d\x64\x7b\x92\xa9"
"\xc0\xd7\xee\x05\x1c\x33\x4a\x61"
"\x78\x8f\xa6\xbd\xd4\xeb\x02\x19"
"\x30\x47\x5e\x75\x8c\xa3\xba\xd1"
"\xe8\xff\x16\x2d\x44\x5b\x72\x89"
"\xa0\xb7\xce\xe5\xfc\x13\x2a\x41"
"\x58\x6f\x86\x9d\xb4\xcb\xe2\xf9"
"\x10\x27\x3e\x55\x6c\x83\x9a\xb1"
"\xc8\xdf\xf6\x0d\x24\x3b\x52\x69"
"\x80\x97\xae\xc5\xdc\xf3\x0a\x21"
"\x38\x4f\x66\x7d\x94\xab\xc2\xd9"
"\xf0\x07\x1e\x35\x4c\x63\x7a\x91"
"\xa8\xbf\xd6\xed\x04\x1b\x32\x49"
"\x60\x77\x8e\xa5\xbc\xd3\xea\x01"
"\x18\x2f\x46\x5d\x74\x8b\xa2\xb9"
"\xd0\xe7\xfe\x15\x2c\x43\x5a\x71"
"\x88\x9f\xb6\xcd\xe4\xfb\x12\x29"
"\x40\x57\x6e\x85\x9c\xb3\xca\xe1"
"\xf8\x0f\x26\x3d\x54\x6b\x82\x99"
"\xb0\xc7\xde\xf5\x0c\x23\x3a\x51"
"\x68\x7f\x96\xad\xc4\xdb\xf2\x09"
"\x20\x37\x4e\x65\x7c\x93\xaa\xc1"
"\xd8\xef\x06\x1d\x34\x4b\x62\x79"
"\x90\xa7\xbe\xd5\xec\x03\x1a\x31"
"\x48\x5f\x76\x8d\xa4\xbb\xd2\xe9"
"\x00\x19\x32\x4b\x64\x7d\x96\xaf"
"\xc8\xe1\xfa\x13\x2c\x45\x5e\x77"
"\x90\xa9\xc2\xdb\xf4\x0d\x26\x3f"
"\x58\x71\x8a\xa3\xbc\xd5\xee\x07"
"\x20\x39\x52\x6b\x84\x9d\xb6\xcf"
"\xe8\x01\x1a\x33\x4c\x65\x7e\x97"
"\xb0\xc9\xe2\xfb\x14\x2d\x46\x5f"
"\x78\x91\xaa\xc3\xdc\xf5\x0e\x27"
"\x40\x59\x72\x8b\xa4\xbd\xd6\xef"
"\x08\x21\x3a\x53\x6c\x85\x9e\xb7"
"\xd0\xe9\x02\x1b\x34\x4d\x66\x7f"
"\x98\xb1\xca\xe3\xfc\x15\x2e\x47"
"\x60\x79\x92\xab\xc4\xdd\xf6\x0f"
"\x28\x41\x5a\x73\x8c\xa5\xbe\xd7"
"\xf0\x09\x22\x3b\x54\x6d\x86\x9f"
"\xb8\xd1\xea\x03\x1c\x35\x4e\x67"
"\x80\x99\xb2\xcb\xe4\xfd\x16\x2f"
"\x48\x61\x7a\x93\xac\xc5\xde\xf7"
"\x10\x29\x42\x5b\x74\x8d\xa6\xbf"
"\xd8\xf1\x0a\x23\x3c\x55\x6e\x87"
"\xa0\xb9\xd2\xeb\x04\x1d\x36\x4f"
"\x68\x81\x9a\xb3\xcc\xe5\xfe\x17"
"\x30\x49\x62\x7b\x94\xad\xc6\xdf"
"\xf8\x11\x2a\x43\x5c\x75\x8e\xa7"
"\xc0\xd9\xf2\x0b\x24\x3d\x56\x6f"
"\x88\xa1\xba\xd3\xec\x05\x1e\x37"
"\x50\x69\x82\x9b\xb4\xcd\xe6\xff"
"\x18\x31\x4a\x63\x7c\x95\xae\xc7"
"\xe0\xf9\x12\x2b\x44\x5d\x76\x8f"
"\xa8\xc1\xda\xf3\x0c\x25\x3e\x57"
"\x70\x89\xa2\xbb\xd4\xed\x06\x1f"
"\x38\x51\x6a\x83\x9c\xb5\xce\xe7"
"\x00\x1b\x36\x51\x6c\x87\xa2\xbd"
"\xd8\xf3\x0e\x29\x44\x5f\x7a\x95"
"\xb0\xcb\xe6\x01\x1c\x37\x52\x6d"
"\x88\xa3\xbe\xd9\xf4\x0f\x2a\x45"
"\x60\x7b\x96\xb1\xcc\xe7\x02\x1d"
"\x38\x53\x6e\x89\xa4\xbf\xda\xf5"
"\x10\x2b\x46\x61\x7c\x97\xb2\xcd"
"\xe8\x03\x1e\x39\x54\x6f\x8a\xa5"
"\xc0\xdb\xf6\x11\x2c\x47\x62\x7d"
"\x98\xb3\xce\xe9\x04\x1f\x3a\x55"
"\x70\x8b\xa6\xc1\xdc\xf7\x12\x2d"
"\x48\x63\x7e\x99\xb4\xcf\xea\x05"
"\x20\x3b\x56\x71\x8c\xa7\xc2\xdd"
"\xf8\x13\x2e\x49\x64\x7f\x9a\xb5"
"\xd0\xeb\x06\x21\x3c\x57\x72\x8d"
"\xa8\xc3\xde\xf9\x14\x2f\x4a\x65"
"\x80\x9b\xb6\xd1\xec\x07\x22\x3d"
"\x58\x73\x8e\xa9\xc4\xdf\xfa\x15"
"\x30\x4b\x66\x81\x9c\xb7\xd2\xed"
"\x08\x23\x3e\x59\x74\x8f\xaa\xc5"
"\xe0\xfb\x16\x31\x4c\x67\x82\x9d"
"\xb8\xd3\xee\x09\x24\x3f\x5a\x75"
"\x90\xab\xc6\xe1\xfc\x17\x32\x4d"
"\x68\x83\x9e\xb9\xd4\xef\x0a\x25"
"\x40\x5b\x76\x91\xac\xc7\xe2\xfd"
"\x18\x33\x4e\x69\x84\x9f\xba\xd5"
"\xf0\x0b\x26\x41\x5c\x77\x92\xad"
"\xc8\xe3\xfe\x19\x34\x4f\x6a\x85"
"\xa0\xbb\xd6\xf1\x0c\x27\x42\x5d"
"\x78\x93\xae\xc9\xe4\xff\x1a\x35"
"\x50\x6b\x86\xa1\xbc\xd7\xf2\x0d"
"\x28\x43\x5e\x79\x94\xaf\xca\xe5"
"\x00\x1d\x3a\x57\x74\x91\xae\xcb"
"\xe8\x05\x22\x3f\x5c\x79\x96\xb3"
"\xd0\xed\x0a\x27\x44\x61\x7e\x9b"
"\xb8\xd5\xf2\x0f\x2c\x49\x66\x83"
"\xa0\xbd\xda\xf7\x14\x31\x4e\x6b"
"\x88\xa5\xc2\xdf\xfc\x19\x36\x53"
"\x70\x8d\xaa\xc7\xe4\x01\x1e\x3b"
"\x58\x75\x92\xaf\xcc\xe9\x06\x23"
"\x40\x5d\x7a\x97\xb4\xd1\xee\x0b"
"\x28\x45\x62\x7f\x9c\xb9\xd6\xf3"
"\x10\x2d\x4a\x67\x84\xa1\xbe\xdb"
"\xf8\x15\x32\x4f\x6c\x89\xa6\xc3"
"\xe0\xfd\x1a\x37\x54\x71\x8e\xab"
"\xc8\xe5\x02\x1f\x3c\x59\x76\x93"
"\xb0\xcd\xea\x07\x24\x41\x5e\x7b"
"\x98\xb5\xd2\xef\x0c\x29\x46\x63"
"\x80\x9d\xba\xd7\xf4\x11\x2e\x4b"
"\x68\x85\xa2\xbf\xdc\xf9\x16\x33"
"\x50\x6d\x8a\xa7\xc4\xe1\xfe\x1b"
"\x38\x55\x72\x8f\xac\xc9\xe6\x03"
"\x20\x3d\x5a\x77\x94\xb1\xce\xeb"
"\x08\x25\x42\x5f\x7c\x99\xb6\xd3"
"\xf0\x0d\x2a\x47\x64\x81\x9e\xbb"
"\xd8\xf5\x12\x2f\x4c\x69\x86\xa3"
"\xc0\xdd\xfa\x17\x34\x51\x6e\x8b"
"\xa8\xc5\xe2\xff\x1c\x39\x56\x73"
"\x90\xad\xca\xe7\x04\x21\x3e\x5b"
"\x78\x95\xb2\xcf\xec\x09\x26\x43"
"\x60\x7d\x9a\xb7\xd4\xf1\x0e\x2b"
"\x48\x65\x82\x9f\xbc\xd9\xf6\x13"
"\x30\x4d\x6a\x87\xa4\xc1\xde\xfb"
"\x18\x35\x52\x6f\x8c\xa9\xc6\xe3"
"\x00\x1f\x3e\x5d\x7c\x9b\xba\xd9"
"\xf8\x17\x36\x55\x74\x93\xb2\xd1"
"\xf0\x0f\x2e\x4d\x6c\x8b\xaa\xc9"
"\xe8\x07\x26\x45\x64\x83\xa2\xc1"
"\xe0\xff\x1e\x3d\x5c\x7b\x9a\xb9"
"\xd8\xf7\x16\x35\x54\x73\x92\xb1"
"\xd0\xef\x0e\x2d\x4c\x6b\x8a\xa9"
"\xc8\xe7\x06\x25\x44\x63\x82\xa1"
"\xc0\xdf\xfe\x1d\x3c\x5b\x7a\x99"
"\xb8\xd7\xf6\x15\x34\x53\x72\x91"
"\xb0\xcf\xee\x0d\x2c\x4b\x6a\x89"
"\xa8\xc7\xe6\x05\x24\x43\x62\x81"
"\xa0\xbf\xde\xfd\x1c\x3b\x5a\x79"
"\x98\xb7\xd6\xf5\x14\x33\x52\x71"
"\x90\xaf\xce\xed\x0c\x2b\x4a\x69"
"\x88\xa7\xc6\xe5\x04\x23\x42\x61"
"\x80\x9f\xbe\xdd\xfc\x1b\x3a\x59"
"\x78\x97\xb6\xd5\xf4\x13\x32\x51"
"\x70\x8f\xae\xcd\xec\x0b\x2a\x49"
"\x68\x87\xa6\xc5\xe4\x03\x22\x41"
"\x60\x7f\x9e\xbd\xdc\xfb\x1a\x39"
"\x58\x77\x96\xb5\xd4\xf3\x12\x31"
"\x50\x6f\x8e\xad\xcc\xeb\x0a\x29"
"\x48\x67\x86\xa5\xc4\xe3\x02\x21"
"\x40\x5f\x7e\x9d\xbc\xdb\xfa\x19"
"\x38\x57\x76\x95\xb4\xd3\xf2\x11"
"\x30\x4f\x6e\x8d\xac\xcb\xea\x09"
"\x28\x47\x66\x85\xa4\xc3\xe2\x01"
"\x20\x3f\x5e\x7d\x9c\xbb\xda\xf9"
"\x18\x37\x56\x75\x94\xb3\xd2\xf1"
"\x10\x2f\x4e\x6d\x8c\xab\xca\xe9"
"\x08\x27\x46\x65\x84\xa3\xc2\xe1"
"\x00\x21\x42\x63",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext =
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xb5\x81\xf5\x64\x18\x73\xe3\xf0"
"\x4c\x13\xf2\x77\x18\x60\x65\x5e"
"\x29\x01\xce\x98\x55\x53\xf9\x0c"
"\x2a\x08\xd5\x09\xb3\x57\x55\x56"
"\xc5\xe9\x56\x90\xcb\x6a\xa3\xc0"
"\xff\xc4\x79\xb4\xd2\x97\x5d\xc4"
"\x43\xd1\xfe\x94\x7b\x88\x06\x5a"
"\xb2\x9e\x2c\xfc\x44\x03\xb7\x90"
"\xa0\xc1\xba\x6a\x33\xb8\xc7\xb2"
"\x9d\xe1\x12\x4f\xc0\x64\xd4\x01"
"\xfe\x8c\x7a\x66\xf7\xe6\x5a\x91"
"\xbb\xde\x56\x86\xab\x65\x21\x30"
"\x00\x84\x65\x24\xa5\x7d\x85\xb4"
"\xe3\x17\xed\x3a\xb7\x6f\xb4\x0b"
"\x0b\xaf\x15\xae\x5a\x8f\xf2\x0c"
"\x2f\x27\xf4\x09\xd8\xd2\x96\xb7"
"\x71\xf2\xc5\x99\x4d\x7e\x7f\x75"
"\x77\x89\x30\x8b\x59\xdb\xa2\xb2"
"\xa0\xf3\x19\x39\x2b\xc5\x7e\x3f"
"\x4f\xd9\xd3\x56\x28\x97\x44\xdc"
"\xc0\x8b\x77\x24\xd9\x52\xe7\xc5"
"\xaf\xf6\x7d\x59\xb2\x44\x05\x1d"
"\xb1\xb0\x11\xa5\x0f\xec\x33\xe1"
"\x6d\x1b\x4e\x1f\xff\x57\x91\xb4"
"\x5b\x9a\x96\xc5\x53\xbc\xae\x20"
"\x3c\xbb\x14\xe2\xe8\x22\x33\xc1"
"\x5e\x76\x9e\x46\x99\xf6\x2a\x15"
"\xc6\x97\x02\xa0\x66\x43\xd1\xa6"
"\x31\xa6\x9f\xfb\xf4\xd3\x69\xe5"
"\xcd\x76\x95\xb8\x7a\x82\x7f\x21"
"\x45\xff\x3f\xce\x55\xf6\x95\x10"
"\x08\x77\x10\x43\xc6\xf3\x09\xe5"
"\x68\xe7\x3c\xad\x00\x52\x45\x0d"
"\xfe\x2d\xc6\xc2\x94\x8c\x12\x1d"
"\xe6\x25\xae\x98\x12\x8e\x19\x9c"
"\x81\x68\xb1\x11\xf6\x69\xda\xe3"
"\x62\x08\x18\x7a\x25\x49\x28\xac"
"\xba\x71\x12\x0b\xe4\xa2\xe5\xc7"
"\x5d\x8e\xec\x49\x40\x21\xbf\x5a"
"\x98\xf3\x02\x68\x55\x03\x7f\x8a"
"\xe5\x94\x0c\x32\x5c\x07\x82\x63"
"\xaf\x6f\x91\x40\x84\x8e\x52\x25"
"\xd0\xb0\x29\x53\x05\xe2\x50\x7a"
"\x34\xeb\xc9\x46\x20\xa8\x3d\xde"
"\x7f\x16\x5f\x36\xc5\x2e\xdc\xd1"
"\x15\x47\xc7\x50\x40\x6d\x91\xc5"
"\xe7\x93\x95\x1a\xd3\x57\xbc\x52"
"\x33\xee\x14\x19\x22\x52\x89\xa7"
"\x4a\x25\x56\x77\x4b\xca\xcf\x0a"
"\xe1\xf5\x35\x85\x30\x7e\x59\x4a"
"\xbd\x14\x5b\xdf\xe3\x46\xcb\xac"
"\x1f\x6c\x96\x0e\xf4\x81\xd1\x99"
"\xca\x88\x63\x3d\x02\x58\x6b\xa9"
"\xe5\x9f\xb3\x00\xb2\x54\xc6\x74"
"\x1c\xbf\x46\xab\x97\xcc\xf8\x54"
"\x04\x07\x08\x52\xe6\xc0\xda\x93"
"\x74\x7d\x93\x99\x5d\x78\x68\xa6"
"\x2e\x6b\xd3\x6a\x69\xcc\x12\x6b"
"\xd4\xc7\xa5\xc6\xe7\xf6\x03\x04"
"\x5d\xcd\x61\x5e\x17\x40\xdc\xd1"
"\x5c\xf5\x08\xdf\x5c\x90\x85\xa4"
"\xaf\xf6\x78\xbb\x0d\xf1\xf4\xa4"
"\x54\x26\x72\x9e\x61\xfa\x86\xcf"
"\xe8\x9e\xa1\xe0\xc7\x48\x23\xae"
"\x5a\x90\xae\x75\x0a\x74\x18\x89"
"\x05\xb1\x92\xb2\x7f\xd0\x1b\xa6"
"\x62\x07\x25\x01\xc7\xc2\x4f\xf9"
"\xe8\xfe\x63\x95\x80\x07\xb4\x26"
"\xcc\xd1\x26\xb6\xc4\x3f\x9e\xcb"
"\x8e\x3b\x2e\x44\x16\xd3\x10\x9a"
"\x95\x08\xeb\xc8\xcb\xeb\xbf\x6f"
"\x0b\xcd\x1f\xc8\xca\x86\xaa\xec"
"\x33\xe6\x69\xf4\x45\x25\x86\x3a"
"\x22\x94\x4f\x00\x23\x6a\x44\xc2"
"\x49\x97\x33\xab\x36\x14\x0a\x70"
"\x24\xc3\xbe\x04\x3b\x79\xa0\xf9"
"\xb8\xe7\x76\x29\x22\x83\xd7\xf2"
"\x94\xf4\x41\x49\xba\x5f\x7b\x07"
"\xb5\xfb\xdb\x03\x1a\x9f\xb6\x4c"
"\xc2\x2e\x37\x40\x49\xc3\x38\x16"
"\xe2\x4f\x77\x82\xb0\x68\x4c\x71"
"\x1d\x57\x61\x9c\xd9\x4e\x54\x99"
"\x47\x13\x28\x73\x3c\xbb\x00\x90"
"\xf3\x4d\xc9\x0e\xfd\xe7\xb1\x71"
"\xd3\x15\x79\xbf\xcc\x26\x2f\xbd"
"\xad\x6c\x50\x69\x6c\x3e\x6d\x80"
"\x9a\xea\x78\xaf\x19\xb2\x0d\x4d"
"\xad\x04\x07\xae\x22\x90\x4a\x93"
"\x32\x0e\x36\x9b\x1b\x46\xba\x3b"
"\xb4\xac\xc6\xd1\xa2\x31\x53\x3b"
"\x2a\x3d\x45\xfe\x03\x61\x10\x85"
"\x17\x69\xa6\x78\xcc\x6c\x87\x49"
"\x53\xf9\x80\x10\xde\x80\xa2\x41"
"\x6a\xc3\x32\x02\xad\x6d\x3c\x56"
"\x00\x71\x51\x06\xa7\xbd\xfb\xef"
"\x3c\xb5\x9f\xfc\x48\x7d\x53\x7c"
"\x66\xb0\x49\x23\xc4\x47\x10\x0e"
"\xe5\x6c\x74\x13\xe6\xc5\x3f\xaa"
"\xde\xff\x07\x44\xdd\x56\x1b\xad"
"\x09\x77\xfb\x5b\x12\xb8\x0d\x38"
"\x17\x37\x35\x7b\x9b\xbc\xfe\xd4"
"\x7e\x8b\xda\x7e\x5b\x04\xa7\x22"
"\xa7\x31\xa1\x20\x86\xc7\x1b\x99"
"\xdb\xd1\x89\xf4\x94\xa3\x53\x69"
"\x8d\xe7\xe8\x74\x11\x8d\x74\xd6"
"\x07\x37\x91\x9f\xfd\x67\x50\x3a"
"\xc9\xe1\xf4\x36\xd5\xa0\x47\xd1"
"\xf9\xe5\x39\xa3\x31\xac\x07\x36"
"\x23\xf8\x66\x18\x14\x28\x34\x0f"
"\xb8\xd0\xe7\x29\xb3\x04\x4b\x55"
"\x01\x41\xb2\x75\x8d\xcb\x96\x85"
"\x3a\xfb\xab\x2b\x9e\xfa\x58\x20"
"\x44\x1f\xc0\x14\x22\x75\x61\xe8"
"\xaa\x19\xcf\xf1\x82\x56\xf4\xd7"
"\x78\x7b\x3d\x5f\xb3\x9e\x0b\x8a"
"\x57\x50\xdb\x17\x41\x65\x4d\xa3"
"\x02\xc9\x9c\x9c\x53\xfb\x39\x39"
"\x9b\x1d\x72\x24\xda\xb7\x39\xbe"
"\x13\x3b\xfa\x29\xda\x9e\x54\x64"
"\x6e\xba\xd8\xa1\xcb\xb3\x36\xfa"
"\xcb\x47\x85\xe9\x61\x38\xbc\xbe"
"\xc5\x00\x38\x2a\x54\xf7\xc4\xb9"
"\xb3\xd3\x7b\xa0\xa0\xf8\x72\x7f"
"\x8c\x8e\x82\x0e\xc6\x1c\x75\x9d"
"\xca\x8e\x61\x87\xde\xad\x80\xd2"
"\xf5\xf9\x80\xef\x15\x75\xaf\xf5"
"\x80\xfb\xff\x6d\x1e\x25\xb7\x40"
"\x61\x6a\x39\x5a\x6a\xb5\x31\xab"
"\x97\x8a\x19\x89\x44\x40\xc0\xa6"
"\xb4\x4e\x30\x32\x7b\x13\xe7\x67"
"\xa9\x8b\x57\x04\xc2\x01\xa6\xf4"
"\x28\x99\xad\x2c\x76\xa3\x78\xc2"
"\x4a\xe6\xca\x5c\x50\x6a\xc1\xb0"
"\x62\x4b\x10\x8e\x7c\x17\x43\xb3"
"\x17\x66\x1c\x3e\x8d\x69\xf0\x5a"
"\x71\xf5\x97\xdc\xd1\x45\xdd\x28"
"\xf3\x5d\xdf\x53\x7b\x11\xe5\xbc"
"\x4c\xdb\x1b\x51\x6b\xe9\xfb\x3d"
"\xc1\xc3\x2c\xb9\x71\xf5\xb6\xb2"
"\x13\x36\x79\x80\x53\xe8\xd3\xa6"
"\x0a\xaf\xfd\x56\x97\xf7\x40\x8e"
"\x45\xce\xf8\xb0\x9e\x5c\x33\x82"
"\xb0\x44\x56\xfc\x05\x09\xe9\x2a"
"\xac\x26\x80\x14\x1d\xc8\x3a\x35"
"\x4c\x82\x97\xfd\x76\xb7\xa9\x0a"
"\x35\x58\x79\x8e\x0f\x66\xea\xaf"
"\x51\x6c\x09\xa9\x6e\x9b\xcb\x9a"
"\x31\x47\xa0\x2f\x7c\x71\xb4\x4a"
"\x11\xaa\x8c\x66\xc5\x64\xe6\x3a"
"\x54\xda\x24\x6a\xc4\x41\x65\x46"
"\x82\xa0\x0a\x0f\x5f\xfb\x25\xd0"
"\x2c\x91\xa7\xee\xc4\x81\x07\x86"
"\x75\x5e\x33\x69\x97\xe4\x2c\xa8"
"\x9d\x9f\x0b\x6a\xbe\xad\x98\xda"
"\x6d\x94\x41\xda\x2c\x1e\x89\xc4"
"\xc2\xaf\x1e\x00\x05\x0b\x83\x60"
"\xbd\x43\xea\x15\x23\x7f\xb9\xac"
"\xee\x4f\x2c\xaf\x2a\xf3\xdf\xd0"
"\xf3\x19\x31\xbb\x4a\x74\x84\x17"
"\x52\x32\x2c\x7d\x61\xe4\xcb\xeb"
"\x80\x38\x15\x52\xcb\x6f\xea\xe5"
"\x73\x9c\xd9\x24\x69\xc6\x95\x32"
"\x21\xc8\x11\xe4\xdc\x36\xd7\x93"
"\x38\x66\xfb\xb2\x7f\x3a\xb9\xaf"
"\x31\xdd\x93\x75\x78\x8a\x2c\x94"
"\x87\x1a\x58\xec\x9e\x7d\x4d\xba"
"\xe1\xe5\x4d\xfc\xbc\xa4\x2a\x14"
"\xef\xcc\xa7\xec\xab\x43\x09\x18"
"\xd3\xab\x68\xd1\x07\x99\x44\x47"
"\xd6\x83\x85\x3b\x30\xea\xa9\x6b"
"\x63\xea\xc4\x07\xfb\x43\x2f\xa4"
"\xaa\xb0\xab\x03\x89\xce\x3f\x8c"
"\x02\x7c\x86\x54\xbc\x88\xaf\x75"
"\xd2\xdc\x63\x17\xd3\x26\xf6\x96"
"\xa9\x3c\xf1\x61\x8c\x11\x18\xcc"
"\xd6\xea\x5b\xe2\xcd\xf0\xf1\xb2"
"\xe5\x35\x90\x1f\x85\x4c\x76\x5b"
"\x66\xce\x44\xa4\x32\x9f\xe6\x7b"
"\x71\x6e\x9f\x58\x15\x67\x72\x87"
"\x64\x8e\x3a\x44\x45\xd4\x76\xfa"
"\xc2\xf6\xef\x85\x05\x18\x7a\x9b"
"\xba\x41\x54\xac\xf0\xfc\x59\x12"
"\x3f\xdf\xa0\xe5\x8a\x65\xfd\x3a"
"\x62\x8d\x83\x2c\x03\xbe\x05\x76"
"\x2e\x53\x49\x97\x94\x33\xae\x40"
"\x81\x15\xdb\x6e\xad\xaa\xf5\x4b"
"\xe3\x98\x70\xdf\xe0\x7c\xcd\xdb"
"\x02\xd4\x7d\x2f\xc1\xe6\xb4\xf3"
"\xd7\x0d\x7a\xd9\x23\x9e\x87\x2d"
"\xce\x87\xad\xcc\x72\x05\x00\x29"
"\xdc\x73\x7f\x64\xc1\x15\x0e\xc2"
"\xdf\xa7\x5f\xeb\x41\xa1\xcd\xef"
"\x5c\x50\x79\x2a\x56\x56\x71\x8c"
"\xac\xc0\x79\x50\x69\xca\x59\x32"
"\x65\xf2\x54\xe4\x52\x38\x76\xd1"
"\x5e\xde\x26\x9e\xfb\x75\x2e\x11"
"\xb5\x10\xf4\x17\x73\xf5\x89\xc7"
"\x4f\x43\x5c\x8e\x7c\xb9\x05\x52"
"\x24\x40\x99\xfe\x9b\x85\x0b\x6c"
"\x22\x3e\x8b\xae\x86\xa1\xd2\x79"
"\x05\x68\x6b\xab\xe3\x41\x49\xed"
"\x15\xa1\x8d\x40\x2d\x61\xdf\x1a"
"\x59\xc9\x26\x8b\xef\x30\x4c\x88"
"\x4b\x10\xf8\x8d\xa6\x92\x9f\x4b"
"\xf3\xc4\x53\x0b\x89\x5d\x28\x92"
"\xcf\x78\xb2\xc0\x5d\xed\x7e\xfc"
"\xc0\x12\x23\x5f\x5a\x78\x86\x43"
"\x6e\x27\xf7\x5a\xa7\x6a\xed\x19"
"\x04\xf0\xb3\x12\xd1\xbd\x0e\x89"
"\x6e\xbc\x96\xa8\xd8\x49\x39\x9f"
"\x7e\x67\xf0\x2e\x3e\x01\xa9\xba"
"\xec\x8b\x62\x8e\xcb\x4a\x70\x43"
"\xc7\xc2\xc4\xca\x82\x03\x73\xe9"
"\x11\xdf\xcf\x54\xea\xc9\xb0\x95"
"\x51\xc0\x13\x3d\x92\x05\xfa\xf4"
"\xa9\x34\xc8\xce\x6c\x3d\x54\xcc"
"\xc4\xaf\xf1\xdc\x11\x44\x26\xa2"
"\xaf\xf1\x85\x75\x7d\x03\x61\x68"
"\x4e\x78\xc6\x92\x7d\x86\x7d\x77"
"\xdc\x71\x72\xdb\xc6\xae\xa1\xcb"
"\x70\x9a\x0b\x19\xbe\x4a\x6c\x2a"
"\xe2\xba\x6c\x64\x9a\x13\x28\xdf"
"\x85\x75\xe6\x43\xf6\x87\x08\x68"
"\x6e\xba\x6e\x79\x9f\x04\xbc\x23"
"\x50\xf6\x33\x5c\x1f\x24\x25\xbe"
"\x33\x47\x80\x45\x56\xa3\xa7\xd7"
"\x7a\xb1\x34\x0b\x90\x3c\x9c\xad"
"\x44\x5f\x9e\x0e\x9d\xd4\xbd\x93"
"\x5e\xfa\x3c\xe0\xb0\xd9\xed\xf3"
"\xd6\x2e\xff\x24\xd8\x71\x6c\xed"
"\xaf\x55\xeb\x22\xac\x93\x68\x32"
"\x05\x5b\x47\xdd\xc6\x4a\xcb\xc7"
"\x10\xe1\x3c\x92\x1a\xf3\x23\x78"
"\x2b\xa1\xd2\x80\xf4\x12\xb1\x20"
"\x8f\xff\x26\x35\xdd\xfb\xc7\x4e"
"\x78\xf1\x2d\x50\x12\x77\xa8\x60"
"\x7c\x0f\xf5\x16\x2f\x63\x70\x2a"
"\xc0\x96\x80\x4e\x0a\xb4\x93\x35"
"\x5d\x1d\x3f\x56\xf7\x2f\xbb\x90"
"\x11\x16\x8f\xa2\xec\x47\xbe\xac"
"\x56\x01\x26\x56\xb1\x8c\xb2\x10"
"\xf9\x1a\xca\xf5\xd1\xb7\x39\x20"
"\x63\xf1\x69\x20\x4f\x13\x12\x1f"
"\x5b\x65\xfc\x98\xf7\xc4\x7a\xbe"
"\xf7\x26\x4d\x2b\x84\x7b\x42\xad"
"\xd8\x7a\x0a\xb4\xd8\x74\xbf\xc1"
"\xf0\x6e\xb4\x29\xa3\xbb\xca\x46"
"\x67\x70\x6a\x2d\xce\x0e\xa2\x8a"
"\xa9\x87\xbf\x05\xc4\xc1\x04\xa3"
"\xab\xd4\x45\x43\x8c\xb6\x02\xb0"
"\x41\xc8\xfc\x44\x3d\x59\xaa\x2e"
"\x44\x21\x2a\x8d\x88\x9d\x57\xf4"
"\xa0\x02\x77\xb8\xa6\xa0\xe6\x75"
"\x5c\x82\x65\x3e\x03\x5c\x29\x8f"
"\x38\x55\xab\x33\x26\xef\x9f\x43"
"\x52\xfd\x68\xaf\x36\xb4\xbb\x9a"
"\x58\x09\x09\x1b\xc3\x65\x46\x46"
"\x1d\xa7\x94\x18\x23\x50\x2c\xca"
"\x2c\x55\x19\x97\x01\x9d\x93\x3b"
"\x63\x86\xf2\x03\x67\x45\xd2\x72"
"\x28\x52\x6c\xf4\xe3\x1c\xb5\x11"
"\x13\xf1\xeb\x21\xc7\xd9\x56\x82"
"\x2b\x82\x39\xbd\x69\x54\xed\x62"
"\xc3\xe2\xde\x73\xd4\x6a\x12\xae"
"\x13\x21\x7f\x4b\x5b\xfc\xbf\xe8"
"\x2b\xbe\x56\xba\x68\x8b\x9a\xb1"
"\x6e\xfa\xbf\x7e\x5a\x4b\xf1\xac"
"\x98\x65\x85\xd1\x93\x53\xd3\x7b"
"\x09\xdd\x4b\x10\x6d\x84\xb0\x13"
"\x65\xbd\xcf\x52\x09\xc4\x85\xe2"
"\x84\x74\x15\x65\xb7\xf7\x51\xaf"
"\x55\xad\xa4\xd1\x22\x54\x70\x94"
"\xa0\x1c\x90\x41\xfd\x99\xd7\x5a"
"\x31\xef\xaa\x25\xd0\x7f\x4f\xea"
"\x1d\x55\x42\xe5\x49\xb0\xd0\x46"
"\x62\x36\x43\xb2\x82\x15\x75\x50"
"\xa4\x72\xeb\x54\x27\x1f\x8a\xe4"
"\x7d\xe9\x66\xc5\xf1\x53\xa4\xd1"
"\x0c\xeb\xb8\xf8\xbc\xd4\xe2\xe7"
"\xe1\xf8\x4b\xcb\xa9\xa1\xaf\x15"
"\x83\xcb\x72\xd0\x33\x79\x00\x2d"
"\x9f\xd7\xf1\x2e\x1e\x10\xe4\x45"
"\xc0\x75\x3a\x39\xea\x68\xf7\x5d"
"\x1b\x73\x8f\xe9\x8e\x0f\x72\x47"
"\xae\x35\x0a\x31\x7a\x14\x4d\x4a"
"\x6f\x47\xf7\x7e\x91\x6e\x74\x8b"
"\x26\x47\xf9\xc3\xf9\xde\x70\xf5"
"\x61\xab\xa9\x27\x9f\x82\xe4\x9c"
"\x89\x91\x3f\x2e\x6a\xfd\xb5\x49"
"\xe9\xfd\x59\x14\x36\x49\x40\x6d"
"\x32\xd8\x85\x42\xf3\xa5\xdf\x0c"
"\xa8\x27\xd7\x54\xe2\x63\x2f\xf2"
"\x7e\x8b\x8b\xe7\xf1\x9a\x95\x35"
"\x43\xdc\x3a\xe4\xb6\xf4\xd0\xdf"
"\x9c\xcb\x94\xf3\x21\xa0\x77\x50"
"\xe2\xc6\xc4\xc6\x5f\x09\x64\x5b"
"\x92\x90\xd8\xe1\xd1\xed\x4b\x42"
"\xd7\x37\xaf\x65\x3d\x11\x39\xb6"
"\x24\x8a\x60\xae\xd6\x1e\xbf\x0e"
"\x0d\xd7\xdc\x96\x0e\x65\x75\x4e"
"\x29\x06\x9d\xa4\x51\x3a\x10\x63"
"\x8f\x17\x07\xd5\x8e\x3c\xf4\x28"
"\x00\x5a\x5b\x05\x19\xd8\xc0\x6c"
"\xe5\x15\xe4\x9c\x9d\x71\x9d\x5e"
"\x94\x29\x1a\xa7\x80\xfa\x0e\x33"
"\x03\xdd\xb7\x3e\x9a\xa9\x26\x18"
"\x37\xa9\x64\x08\x4d\x94\x5a\x88"
"\xca\x35\xce\x81\x02\xe3\x1f\x1b"
"\x89\x1a\x77\x85\xe3\x41\x6d\x32"
"\x42\x19\x23\x7d\xc8\x73\xee\x25"
"\x85\x0d\xf8\x31\x25\x79\x1b\x6f"
"\x79\x25\xd2\xd8\xd4\x23\xfd\xf7"
"\x82\x36\x6a\x0c\x46\x22\x15\xe9"
"\xff\x72\x41\x91\x91\x7d\x3a\xb7"
"\xdd\x65\x99\x70\xf6\x8d\x84\xf8"
"\x67\x15\x20\x11\xd6\xb2\x55\x7b"
"\xdb\x87\xee\xef\x55\x89\x2a\x59"
"\x2b\x07\x8f\x43\x8a\x59\x3c\x01"
"\x8b\x65\x54\xa1\x66\xd5\x38\xbd"
"\xc6\x30\xa9\xcc\x49\xb6\xa8\x1b"
"\xb8\xc0\x0e\xe3\x45\x28\xe2\xff"
"\x41\x9f\x7e\x7c\xd1\xae\x9e\x25"
"\x3f\x4c\x7c\x7c\xf4\xa8\x26\x4d"
"\x5c\xfd\x4b\x27\x18\xf9\x61\x76"
"\x48\xba\x0c\x6b\xa9\x4d\xfc\xf5"
"\x3b\x35\x7e\x2f\x4a\xa9\xc2\x9a"
"\xae\xab\x86\x09\x89\xc9\xc2\x40"
"\x39\x2c\x81\xb3\xb8\x17\x67\xc2"
"\x0d\x32\x4a\x3a\x67\x81\xd7\x1a"
"\x34\x52\xc5\xdb\x0a\xf5\x63\x39"
"\xea\x1f\xe1\x7c\xa1\x9e\xc1\x35"
"\xe3\xb1\x18\x45\x67\xf9\x22\x38"
"\x95\xd9\x34\x34\x86\xc6\x41\x94"
"\x15\xf9\x5b\x41\xa6\x87\x8b\xf8"
"\xd5\xe1\x1b\xe2\x5b\xf3\x86\x10"
"\xff\xe6\xae\x69\x76\xbc\x0d\xb4"
"\x09\x90\x0c\xa2\x65\x0c\xad\x74"
"\xf5\xd7\xff\xda\xc1\xce\x85\xbe"
"\x00\xa7\xff\x4d\x2f\x65\xd3\x8c"
"\x86\x2d\x05\xe8\xed\x3e\x6b\x8b"
"\x0f\x3d\x83\x8c\xf1\x1d\x5b\x96"
"\x2e\xb1\x9c\xc2\x98\xe1\x70\xb9"
"\xba\x5c\x8a\x43\xd6\x34\xa7\x2d"
"\xc9\x92\xae\xf2\xa5\x7b\x05\x49"
"\xa7\x33\x34\x86\xca\xe4\x96\x23"
"\x76\x5b\xf2\xc6\xf1\x51\x28\x42"
"\x7b\xcc\x76\x8f\xfa\xa2\xad\x31"
"\xd4\xd6\x7a\x6d\x25\x25\x54\xe4"
"\x3f\x50\x59\xe1\x5c\x05\xb7\x27"
"\x48\xbf\x07\xec\x1b\x13\xbe\x2b"
"\xa1\x57\x2b\xd5\xab\xd7\xd0\x4c"
"\x1e\xcb\x71\x9b\xc5\x90\x85\xd3"
"\xde\x59\xec\x71\xeb\x89\xbb\xd0"
"\x09\x50\xe1\x16\x3f\xfd\x1c\x34"
"\xc3\x1c\xa1\x10\x77\x53\x98\xef"
"\xf2\xfd\xa5\x01\x59\xc2\x9b\x26"
"\xc7\x42\xd9\x49\xda\x58\x2b\x6e"
"\x9f\x53\x19\x76\x7e\xd9\xc9\x0e"
"\x68\xc8\x7f\x51\x22\x42\xef\x49"
"\xa4\x55\xb6\x36\xac\x09\xc7\x31"
"\x88\x15\x4b\x2e\x8f\x3a\x08\xf7"
"\xd8\xf7\xa8\xc5\xa9\x33\xa6\x45"
"\xe4\xc4\x94\x76\xf3\x0d\x8f\x7e"
"\xc8\xf6\xbc\x23\x0a\xb6\x4c\xd3"
"\x6a\xcd\x36\xc2\x90\x5c\x5c\x3c"
"\x65\x7b\xc2\xd6\xcc\xe6\x0d\x87"
"\x73\x2e\x71\x79\x16\x06\x63\x28"
"\x09\x15\xd8\x89\x38\x38\x3d\xb5"
"\x42\x1c\x08\x24\xf7\x2a\xd2\x9d"
"\xc8\xca\xef\xf9\x27\xd8\x07\x86"
"\xf7\x43\x0b\x55\x15\x3f\x9f\x83"
"\xef\xdc\x49\x9d\x2a\xc1\x54\x62"
"\xbd\x9b\x66\x55\x9f\xb7\x12\xf3"
"\x1b\x4d\x9d\x2a\x5c\xed\x87\x75"
"\x87\x26\xec\x61\x2c\xb4\x0f\x89"
"\xb0\xfb\x2e\x68\x5d\x15\xc7\x8d"
"\x2e\xc0\xd9\xec\xaf\x4f\xd2\x25"
"\x29\xe8\xd2\x26\x2b\x67\xe9\xfc"
"\x2b\xa8\x67\x96\x12\x1f\x5b\x96"
"\xc6\x14\x53\xaf\x44\xea\xd6\xe2"
"\x94\x98\xe4\x12\x93\x4c\x92\xe0"
"\x18\xa5\x8d\x2d\xe4\x71\x3c\x47"
"\x4c\xf7\xe6\x47\x9e\xc0\x68\xdf"
"\xd4\xf5\x5a\x74\xb1\x2b\x29\x03"
"\x19\x07\xaf\x90\x62\x5c\x68\x98"
"\x48\x16\x11\x02\x9d\xee\xb4\x9b"
"\xe5\x42\x7f\x08\xfd\x16\x32\x0b"
"\xd0\xb3\xfa\x2b\xb7\x99\xf9\x29"
"\xcd\x20\x45\x9f\xb3\x1a\x5d\xa2"
"\xaf\x4d\xe0\xbd\x42\x0d\xbc\x74"
"\x99\x9c\x8e\x53\x1a\xb4\x3e\xbd"
"\xa2\x9a\x2d\xf7\xf8\x39\x0f\x67"
"\x63\xfc\x6b\xc0\xaf\xb3\x4b\x4f"
"\x55\xc4\xcf\xa7\xc8\x04\x11\x3e"
"\x14\x32\xbb\x1b\x38\x77\xd6\x7f"
"\x54\x4c\xdf\x75\xf3\x07\x2d\x33"
"\x9b\xa8\x20\xe1\x7b\x12\xb5\xf3"
"\xef\x2f\xce\x72\xe5\x24\x60\xc1"
"\x30\xe2\xab\xa1\x8e\x11\x09\xa8"
"\x21\x33\x44\xfe\x7f\x35\x32\x93"
"\x39\xa7\xad\x8b\x79\x06\xb2\xcb"
"\x4e\xa9\x5f\xc7\xba\x74\x29\xec"
"\x93\xa0\x4e\x54\x93\xc0\xbc\x55"
"\x64\xf0\x48\xe5\x57\x99\xee\x75"
"\xd6\x79\x0f\x66\xb7\xc6\x57\x76"
"\xf7\xb7\xf3\x9c\xc5\x60\xe8\x7f"
"\x83\x76\xd6\x0e\xaa\xe6\x90\x39"
"\x1d\xa6\x32\x6a\x34\xe3\x55\xf8"
"\x58\xa0\x58\x7d\x33\xe0\x22\x39"
"\x44\x64\x87\x86\x5a\x2f\xa7\x7e"
"\x0f\x38\xea\xb0\x30\xcc\x61\xa5"
"\x6a\x32\xae\x1e\xf7\xe9\xd0\xa9"
"\x0c\x32\x4b\xb5\x49\x28\xab\x85"
"\x2f\x8e\x01\x36\x38\x52\xd0\xba"
"\xd6\x02\x78\xf8\x0e\x3e\x9c\x8b"
"\x6b\x45\x99\x3f\x5c\xfe\x58\xf1"
"\x5c\x94\x04\xe1\xf5\x18\x6d\x51"
"\xb2\x5d\x18\x20\xb6\xc2\x9a\x42"
"\x1d\xb3\xab\x3c\xb6\x3a\x13\x03"
"\xb2\x46\x82\x4f\xfc\x64\xbc\x4f"
"\xca\xfa\x9c\xc0\xd5\xa7\xbd\x11"
"\xb7\xe4\x5a\xf6\x6f\x4d\x4d\x54"
"\xea\xa4\x98\x66\xd4\x22\x3b\xd3"
"\x8f\x34\x47\xd9\x7c\xf4\x72\x3b"
"\x4d\x02\x77\xf6\xd6\xdd\x08\x0a"
"\x81\xe1\x86\x89\x3e\x56\x10\x3c"
"\xba\xd7\x81\x8c\x08\xbc\x8b\xe2"
"\x53\xec\xa7\x89\xee\xc8\x56\xb5"
"\x36\x2c\xb2\x03\xba\x99\xdd\x7c"
"\x48\xa0\xb0\xbc\x91\x33\xe9\xa8"
"\xcb\xcd\xcf\x59\x5f\x1f\x15\xe2"
"\x56\xf5\x4e\x01\x35\x27\x45\x77"
"\x47\xc8\xbc\xcb\x7e\x39\xc1\x97"
"\x28\xd3\x84\xfc\x2c\x3e\xc8\xad"
"\x9c\xf8\x8a\x61\x9c\x28\xaa\xc5"
"\x99\x20\x43\x85\x9d\xa5\xe2\x8b"
"\xb8\xae\xeb\xd0\x32\x0d\x52\x78"
"\x09\x56\x3f\xc7\xd8\x7e\x26\xfc"
"\x37\xfb\x6f\x04\xfc\xfa\x92\x10"
"\xac\xf8\x3e\x21\xdc\x8c\x21\x16"
"\x7d\x67\x6e\xf6\xcd\xda\xb6\x98"
"\x23\xab\x23\x3c\xb2\x10\xa0\x53"
"\x5a\x56\x9f\xc5\xd0\xff\xbb\xe4"
"\x98\x3c\x69\x1e\xdb\x38\x8f\x7e"
"\x0f\xd2\x98\x88\x81\x8b\x45\x67"
"\xea\x33\xf1\xeb\xe9\x97\x55\x2e"
"\xd9\xaa\xeb\x5a\xec\xda\xe1\x68"
"\xa8\x9d\x3c\x84\x7c\x05\x3d\x62"
"\x87\x8f\x03\x21\x28\x95\x0c\x89"
"\x25\x22\x4a\xb0\x93\xa9\x50\xa2"
"\x2f\x57\x6e\x18\x42\x19\x54\x0c"
"\x55\x67\xc6\x11\x49\xf4\x5c\xd2"
"\xe9\x3d\xdd\x8b\x48\x71\x21\x00"
"\xc3\x9a\x6c\x85\x74\x28\x83\x4a"
"\x1b\x31\x05\xe1\x06\x92\xe7\xda"
"\x85\x73\x78\x45\x20\x7f\xae\x13"
"\x7c\x33\x06\x22\xf4\x83\xf9\x35"
"\x3f\x6c\x71\xa8\x4e\x48\xbe\x9b"
"\xce\x8a\xba\xda\xbe\x28\x08\xf7"
"\xe2\x14\x8c\x71\xea\x72\xf9\x33"
"\xf2\x88\x3f\xd7\xbb\x69\x6c\x29"
"\x19\xdc\x84\xce\x1f\x12\x4f\xc8"
"\xaf\xa5\x04\xba\x5a\xab\xb0\xd9"
"\x14\x1f\x6c\x68\x98\x39\x89\x7a"
"\xd9\xd8\x2f\xdf\xa8\x47\x4a\x25"
"\xe2\xfb\x33\xf4\x59\x78\xe1\x68"
"\x85\xcf\xfe\x59\x20\xd4\x05\x1d"
"\x80\x99\xae\xbc\xca\xae\x0f\x2f"
"\x65\x43\x34\x8e\x7e\xac\xd3\x93"
"\x2f\xac\x6d\x14\x3d\x02\x07\x70"
"\x9d\xa4\xf3\x1b\x5c\x36\xfc\x01"
"\x73\x34\x85\x0c\x6c\xd6\xf1\xbd"
"\x3f\xdf\xee\xf5\xd9\xba\x56\xef"
"\xf4\x9b\x6b\xee\x9f\x5a\x78\x6d"
"\x32\x19\xf4\xf7\xf8\x4c\x69\x0b"
"\x4b\xbc\xbb\xb7\xf2\x85\xaf\x70"
"\x75\x24\x6c\x54\xa7\x0e\x4d\x1d"
"\x01\xbf\x08\xac\xcf\x7f\x2c\xe3"
"\x14\x89\x5e\x70\x5a\x99\x92\xcd"
"\x01\x84\xc8\xd2\xab\xe5\x4f\x58"
"\xe7\x0f\x2f\x0e\xff\x68\xea\xfd"
"\x15\xb3\x17\xe6\xb0\xe7\x85\xd8"
"\x23\x2e\x05\xc7\xc9\xc4\x46\x1f"
"\xe1\x9e\x49\x20\x23\x24\x4d\x7e"
"\x29\x65\xff\xf4\xb6\xfd\x1a\x85"
"\xc4\x16\xec\xfc\xea\x7b\xd6\x2c"
"\x43\xf8\xb7\xbf\x79\xc0\x85\xcd"
"\xef\xe1\x98\xd3\xa5\xf7\x90\x8c"
"\xe9\x7f\x80\x6b\xd2\xac\x4c\x30"
"\xa7\xc6\x61\x6c\xd2\xf9\x2c\xff"
"\x30\xbc\x22\x81\x7d\x93\x12\xe4"
"\x0a\xcd\xaf\xdd\xe8\xab\x0a\x1e"
"\x13\xa4\x27\xc3\x5f\xf7\x4b\xbb"
"\x37\x09\x4b\x91\x6f\x92\x4f\xaf"
"\x52\xee\xdf\xef\x09\x6f\xf7\x5c"
"\x6e\x12\x17\x72\x63\x57\xc7\xba"
"\x3b\x6b\x38\x32\x73\x1b\x9c\x80"
"\xc1\x7a\xc6\xcf\xcd\x35\xc0\x6b"
"\x31\x1a\x6b\xe9\xd8\x2c\x29\x3f"
"\x96\xfb\xb6\xcd\x13\x91\x3b\xc2"
"\xd2\xa3\x31\x8d\xa4\xcd\x57\xcd"
"\x13\x3d\x64\xfd\x06\xce\xe6\xdc"
"\x0c\x24\x43\x31\x40\x57\xf1\x72"
"\x17\xe3\x3a\x63\x6d\x35\xcf\x5d"
"\x97\x40\x59\xdd\xf7\x3c\x02\xf7"
"\x1c\x7e\x05\xbb\xa9\x0d\x01\xb1"
"\x8e\xc0\x30\xa9\x53\x24\xc9\x89"
"\x84\x6d\xaa\xd0\xcd\x91\xc2\x4d"
"\x91\xb0\x89\xe2\xbf\x83\x44\xaa"
"\x28\x72\x23\xa0\xc2\xad\xad\x1c"
"\xfc\x3f\x09\x7a\x0b\xdc\xc5\x1b"
"\x87\x13\xc6\x5b\x59\x8d\xf2\xc8"
"\xaf\xdf\x11\x95",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 4100,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.np = 2,
.tap = { 4064, 36 },
},
};
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec chacha20_tv_template[] = {
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
{ /* RFC7539 A.2. Test Vector #1 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
.klen = 32,
.iv = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x00\x00\x00\x00\x00\x00\x00\x00"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x76\xb8\xe0\xad\xa0\xf1\x3d\x90"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x40\x5d\x6a\xe5\x53\x86\xbd\x28"
"\xbd\xd2\x19\xb8\xa0\x8d\xed\x1a"
"\xa8\x36\xef\xcc\x8b\x77\x0d\xc7"
"\xda\x41\x59\x7c\x51\x57\x48\x8d"
"\x77\x24\xe0\x3f\xb8\xd8\x4a\x37"
"\x6a\x43\xb8\xf4\x15\x18\xa1\x1c"
"\xc3\x87\xb6\x69\xb2\xee\x65\x86",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 64,
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
}, { /* RFC7539 A.2. Test Vector #2 */
.key = "\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
.klen = 32,
.iv = "\x01\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x41\x6e\x79\x20\x73\x75\x62\x6d"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x69\x73\x73\x69\x6f\x6e\x20\x74"
"\x6f\x20\x74\x68\x65\x20\x49\x45"
"\x54\x46\x20\x69\x6e\x74\x65\x6e"
"\x64\x65\x64\x20\x62\x79\x20\x74"
"\x68\x65\x20\x43\x6f\x6e\x74\x72"
"\x69\x62\x75\x74\x6f\x72\x20\x66"
"\x6f\x72\x20\x70\x75\x62\x6c\x69"
"\x63\x61\x74\x69\x6f\x6e\x20\x61"
"\x73\x20\x61\x6c\x6c\x20\x6f\x72"
"\x20\x70\x61\x72\x74\x20\x6f\x66"
"\x20\x61\x6e\x20\x49\x45\x54\x46"
"\x20\x49\x6e\x74\x65\x72\x6e\x65"
"\x74\x2d\x44\x72\x61\x66\x74\x20"
"\x6f\x72\x20\x52\x46\x43\x20\x61"
"\x6e\x64\x20\x61\x6e\x79\x20\x73"
"\x74\x61\x74\x65\x6d\x65\x6e\x74"
"\x20\x6d\x61\x64\x65\x20\x77\x69"
"\x74\x68\x69\x6e\x20\x74\x68\x65"
"\x20\x63\x6f\x6e\x74\x65\x78\x74"
"\x20\x6f\x66\x20\x61\x6e\x20\x49"
"\x45\x54\x46\x20\x61\x63\x74\x69"
"\x76\x69\x74\x79\x20\x69\x73\x20"
"\x63\x6f\x6e\x73\x69\x64\x65\x72"
"\x65\x64\x20\x61\x6e\x20\x22\x49"
"\x45\x54\x46\x20\x43\x6f\x6e\x74"
"\x72\x69\x62\x75\x74\x69\x6f\x6e"
"\x22\x2e\x20\x53\x75\x63\x68\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x63\x6c\x75"
"\x64\x65\x20\x6f\x72\x61\x6c\x20"
"\x73\x74\x61\x74\x65\x6d\x65\x6e"
"\x74\x73\x20\x69\x6e\x20\x49\x45"
"\x54\x46\x20\x73\x65\x73\x73\x69"
"\x6f\x6e\x73\x2c\x20\x61\x73\x20"
"\x77\x65\x6c\x6c\x20\x61\x73\x20"
"\x77\x72\x69\x74\x74\x65\x6e\x20"
"\x61\x6e\x64\x20\x65\x6c\x65\x63"
"\x74\x72\x6f\x6e\x69\x63\x20\x63"
"\x6f\x6d\x6d\x75\x6e\x69\x63\x61"
"\x74\x69\x6f\x6e\x73\x20\x6d\x61"
"\x64\x65\x20\x61\x74\x20\x61\x6e"
"\x79\x20\x74\x69\x6d\x65\x20\x6f"
"\x72\x20\x70\x6c\x61\x63\x65\x2c"
"\x20\x77\x68\x69\x63\x68\x20\x61"
"\x72\x65\x20\x61\x64\x64\x72\x65"
"\x73\x73\x65\x64\x20\x74\x6f",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\xa3\xfb\xf0\x7d\xf3\xfa\x2f\xde"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x4f\x37\x6c\xa2\x3e\x82\x73\x70"
"\x41\x60\x5d\x9f\x4f\x4f\x57\xbd"
"\x8c\xff\x2c\x1d\x4b\x79\x55\xec"
"\x2a\x97\x94\x8b\xd3\x72\x29\x15"
"\xc8\xf3\xd3\x37\xf7\xd3\x70\x05"
"\x0e\x9e\x96\xd6\x47\xb7\xc3\x9f"
"\x56\xe0\x31\xca\x5e\xb6\x25\x0d"
"\x40\x42\xe0\x27\x85\xec\xec\xfa"
"\x4b\x4b\xb5\xe8\xea\xd0\x44\x0e"
"\x20\xb6\xe8\xdb\x09\xd8\x81\xa7"
"\xc6\x13\x2f\x42\x0e\x52\x79\x50"
"\x42\xbd\xfa\x77\x73\xd8\xa9\x05"
"\x14\x47\xb3\x29\x1c\xe1\x41\x1c"
"\x68\x04\x65\x55\x2a\xa6\xc4\x05"
"\xb7\x76\x4d\x5e\x87\xbe\xa8\x5a"
"\xd0\x0f\x84\x49\xed\x8f\x72\xd0"
"\xd6\x62\xab\x05\x26\x91\xca\x66"
"\x42\x4b\xc8\x6d\x2d\xf8\x0e\xa4"
"\x1f\x43\xab\xf9\x37\xd3\x25\x9d"
"\xc4\xb2\xd0\xdf\xb4\x8a\x6c\x91"
"\x39\xdd\xd7\xf7\x69\x66\xe9\x28"
"\xe6\x35\x55\x3b\xa7\x6c\x5c\x87"
"\x9d\x7b\x35\xd4\x9e\xb2\xe6\x2b"
"\x08\x71\xcd\xac\x63\x89\x39\xe2"
"\x5e\x8a\x1e\x0e\xf9\xd5\x28\x0f"
"\xa8\xca\x32\x8b\x35\x1c\x3c\x76"
"\x59\x89\xcb\xcf\x3d\xaa\x8b\x6c"
"\xcc\x3a\xaf\x9f\x39\x79\xc9\x2b"
"\x37\x20\xfc\x88\xdc\x95\xed\x84"
"\xa1\xbe\x05\x9c\x64\x99\xb9\xfd"
"\xa2\x36\xe7\xe8\x18\xb0\x4b\x0b"
"\xc3\x9c\x1e\x87\x6b\x19\x3b\xfe"
"\x55\x69\x75\x3f\x88\x12\x8c\xc0"
"\x8a\xaa\x9b\x63\xd1\xa1\x6f\x80"
"\xef\x25\x54\xd7\x18\x9c\x41\x1f"
"\x58\x69\xca\x52\xc5\xb8\x3f\xa3"
"\x6f\xf2\x16\xb9\xc1\xd3\x00\x62"
"\xbe\xbc\xfd\x2d\xc5\xbc\xe0\x91"
"\x19\x34\xfd\xa7\x9a\x86\xf6\xe6"
"\x98\xce\xd7\x59\xc3\xff\x9b\x64"
"\x77\x33\x8f\x3d\xa4\xf9\xcd\x85"
"\x14\xea\x99\x82\xcc\xaf\xb3\x41"
"\xb2\x38\x4d\xd9\x02\xf3\xd1\xab"
"\x7a\xc6\x1d\xd2\x9c\x6f\x21\xba"
"\x5b\x86\x2f\x37\x30\xe3\x7c\xfd"
"\xc4\xfd\x80\x6c\x22\xf2\x21",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 375,
crypto: testmgr - add chunked test cases for chacha20 We failed to catch a bug in the chacha20 code after porting it to the skcipher API. We would have caught it if any chunked tests had been defined, so define some now so we will catch future regressions. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-08-14 07:28:15 -06:00
.also_non_np = 1,
.np = 3,
.tap = { 375 - 20, 4, 16 },
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
}, { /* RFC7539 A.2. Test Vector #3 */
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0",
.klen = 32,
.iv = "\x2a\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x02",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x27\x54\x77\x61\x73\x20\x62\x72"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x69\x6c\x6c\x69\x67\x2c\x20\x61"
"\x6e\x64\x20\x74\x68\x65\x20\x73"
"\x6c\x69\x74\x68\x79\x20\x74\x6f"
"\x76\x65\x73\x0a\x44\x69\x64\x20"
"\x67\x79\x72\x65\x20\x61\x6e\x64"
"\x20\x67\x69\x6d\x62\x6c\x65\x20"
"\x69\x6e\x20\x74\x68\x65\x20\x77"
"\x61\x62\x65\x3a\x0a\x41\x6c\x6c"
"\x20\x6d\x69\x6d\x73\x79\x20\x77"
"\x65\x72\x65\x20\x74\x68\x65\x20"
"\x62\x6f\x72\x6f\x67\x6f\x76\x65"
"\x73\x2c\x0a\x41\x6e\x64\x20\x74"
"\x68\x65\x20\x6d\x6f\x6d\x65\x20"
"\x72\x61\x74\x68\x73\x20\x6f\x75"
"\x74\x67\x72\x61\x62\x65\x2e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x62\xe6\x34\x7f\x95\xed\x87\xa4"
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
"\x5f\xfa\xe7\x42\x6f\x27\xa1\xdf"
"\x5f\xb6\x91\x10\x04\x4c\x0d\x73"
"\x11\x8e\xff\xa9\x5b\x01\xe5\xcf"
"\x16\x6d\x3d\xf2\xd7\x21\xca\xf9"
"\xb2\x1e\x5f\xb1\x4c\x61\x68\x71"
"\xfd\x84\xc5\x4f\x9d\x65\xb2\x83"
"\x19\x6c\x7f\xe4\xf6\x05\x53\xeb"
"\xf3\x9c\x64\x02\xc4\x22\x34\xe3"
"\x2a\x35\x6b\x3e\x76\x43\x12\xa6"
"\x1a\x55\x32\x05\x57\x16\xea\xd6"
"\x96\x25\x68\xf8\x7d\x3f\x3f\x77"
"\x04\xc6\xa8\xd1\xbc\xd1\xbf\x4d"
"\x50\xd6\x15\x4b\x6d\xa7\x31\xb1"
"\x87\xb5\x8d\xfd\x72\x8a\xfa\x36"
"\x75\x7a\x79\x7a\xc1\x88\xd1",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 127,
crypto: testmgr - Add a longer ChaCha20 test vector The AVX2 variant of ChaCha20 is used only for messages with >= 512 bytes length. With the existing test vectors, the implementation could not be tested. Due that lack of such a long official test vector, this one is self-generated using chacha20-generic. Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-16 11:14:04 -06:00
}, { /* Self-made test vector for long data */
.key = "\x1c\x92\x40\xa5\xeb\x55\xd3\x8a"
"\xf3\x33\x88\x86\x04\xf6\xb5\xf0"
"\x47\x39\x17\xc1\x40\x2b\x80\x09"
"\x9d\xca\x5c\xbc\x20\x70\x75\xc0",
.klen = 32,
.iv = "\x1c\x00\x00\x00\x00\x00\x00\x00"
"\x00\x00\x00\x00\x00\x00\x00\x01",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\xee\xe0\xdc\x24\x90\x40\xcd"
crypto: testmgr - Add a longer ChaCha20 test vector The AVX2 variant of ChaCha20 is used only for messages with >= 512 bytes length. With the existing test vectors, the implementation could not be tested. Due that lack of such a long official test vector, this one is self-generated using chacha20-generic. Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-16 11:14:04 -06:00
"\xc5\x40\x8f\x47\x05\xbc\xdd\x81"
"\x47\xc6\x8d\xe6\xb1\x8f\xd7\xcb"
"\x09\x0e\x6e\x22\x48\x1f\xbf\xb8"
"\x5c\xf7\x1e\x8a\xc1\x23\xf2\xd4"
"\x19\x4b\x01\x0f\x4e\xa4\x43\xce"
"\x01\xc6\x67\xda\x03\x91\x18\x90"
"\xa5\xa4\x8e\x45\x03\xb3\x2d\xac"
"\x74\x92\xd3\x53\x47\xc8\xdd\x25"
"\x53\x6c\x02\x03\x87\x0d\x11\x0c"
"\x58\xe3\x12\x18\xfd\x2a\x5b\x40"
"\x0c\x30\xf0\xb8\x3f\x43\xce\xae"
"\x65\x3a\x7d\x7c\xf4\x54\xaa\xcc"
"\x33\x97\xc3\x77\xba\xc5\x70\xde"
"\xd7\xd5\x13\xa5\x65\xc4\x5f\x0f"
"\x46\x1a\x0d\x97\xb5\xf3\xbb\x3c"
"\x84\x0f\x2b\xc5\xaa\xea\xf2\x6c"
"\xc9\xb5\x0c\xee\x15\xf3\x7d\xbe"
"\x9f\x7b\x5a\xa6\xae\x4f\x83\xb6"
"\x79\x49\x41\xf4\x58\x18\xcb\x86"
"\x7f\x30\x0e\xf8\x7d\x44\x36\xea"
"\x75\xeb\x88\x84\x40\x3c\xad\x4f"
"\x6f\x31\x6b\xaa\x5d\xe5\xa5\xc5"
"\x21\x66\xe9\xa7\xe3\xb2\x15\x88"
"\x78\xf6\x79\xa1\x59\x47\x12\x4e"
"\x9f\x9f\x64\x1a\xa0\x22\x5b\x08"
"\xbe\x7c\x36\xc2\x2b\x66\x33\x1b"
"\xdd\x60\x71\xf7\x47\x8c\x61\xc3"
"\xda\x8a\x78\x1e\x16\xfa\x1e\x86"
"\x81\xa6\x17\x2a\xa7\xb5\xc2\xe7"
"\xa4\xc7\x42\xf1\xcf\x6a\xca\xb4"
"\x45\xcf\xf3\x93\xf0\xe7\xea\xf6"
"\xf4\xe6\x33\x43\x84\x93\xa5\x67"
"\x9b\x16\x58\x58\x80\x0f\x2b\x5c"
"\x24\x74\x75\x7f\x95\x81\xb7\x30"
"\x7a\x33\xa7\xf7\x94\x87\x32\x27"
"\x10\x5d\x14\x4c\x43\x29\xdd\x26"
"\xbd\x3e\x3c\x0e\xfe\x0e\xa5\x10"
"\xea\x6b\x64\xfd\x73\xc6\xed\xec"
"\xa8\xc9\xbf\xb3\xba\x0b\x4d\x07"
"\x70\xfc\x16\xfd\x79\x1e\xd7\xc5"
"\x49\x4e\x1c\x8b\x8d\x79\x1b\xb1"
"\xec\xca\x60\x09\x4c\x6a\xd5\x09"
"\x49\x46\x00\x88\x22\x8d\xce\xea"
"\xb1\x17\x11\xde\x42\xd2\x23\xc1"
"\x72\x11\xf5\x50\x73\x04\x40\x47"
"\xf9\x5d\xe7\xa7\x26\xb1\x7e\xb0"
"\x3f\x58\xc1\x52\xab\x12\x67\x9d"
"\x3f\x43\x4b\x68\xd4\x9c\x68\x38"
"\x07\x8a\x2d\x3e\xf3\xaf\x6a\x4b"
"\xf9\xe5\x31\x69\x22\xf9\xa6\x69"
"\xc6\x9c\x96\x9a\x12\x35\x95\x1d"
"\x95\xd5\xdd\xbe\xbf\x93\x53\x24"
"\xfd\xeb\xc2\x0a\x64\xb0\x77\x00"
"\x6f\x88\xc4\x37\x18\x69\x7c\xd7"
"\x41\x92\x55\x4c\x03\xa1\x9a\x4b"
"\x15\xe5\xdf\x7f\x37\x33\x72\xc1"
"\x8b\x10\x67\xa3\x01\x57\x94\x25"
"\x7b\x38\x71\x7e\xdd\x1e\xcc\x73"
"\x55\xd2\x8e\xeb\x07\xdd\xf1\xda"
"\x58\xb1\x47\x90\xfe\x42\x21\x72"
"\xa3\x54\x7a\xa0\x40\xec\x9f\xdd"
"\xc6\x84\x6e\xca\xae\xe3\x68\xb4"
"\x9d\xe4\x78\xff\x57\xf2\xf8\x1b"
"\x03\xa1\x31\xd9\xde\x8d\xf5\x22"
"\x9c\xdd\x20\xa4\x1e\x27\xb1\x76"
"\x4f\x44\x55\xe2\x9b\xa1\x9c\xfe"
"\x54\xf7\x27\x1b\xf4\xde\x02\xf5"
"\x1b\x55\x48\x5c\xdc\x21\x4b\x9e"
"\x4b\x6e\xed\x46\x23\xdc\x65\xb2"
"\xcf\x79\x5f\x28\xe0\x9e\x8b\xe7"
"\x4c\x9d\x8a\xff\xc1\xa6\x28\xb8"
"\x65\x69\x8a\x45\x29\xef\x74\x85"
"\xde\x79\xc7\x08\xae\x30\xb0\xf4"
"\xa3\x1d\x51\x41\xab\xce\xcb\xf6"
"\xb5\xd8\x6d\xe0\x85\xe1\x98\xb3"
"\x43\xbb\x86\x83\x0a\xa0\xf5\xb7"
"\x04\x0b\xfa\x71\x1f\xb0\xf6\xd9"
"\x13\x00\x15\xf0\xc7\xeb\x0d\x5a"
"\x9f\xd7\xb9\x6c\x65\x14\x22\x45"
"\x6e\x45\x32\x3e\x7e\x60\x1a\x12"
"\x97\x82\x14\xfb\xaa\x04\x22\xfa"
"\xa0\xe5\x7e\x8c\x78\x02\x48\x5d"
"\x78\x33\x5a\x7c\xad\xdb\x29\xce"
"\xbb\x8b\x61\xa4\xb7\x42\xe2\xac"
"\x8b\x1a\xd9\x2f\x0b\x8b\x62\x21"
"\x83\x35\x7e\xad\x73\xc2\xb5\x6c"
"\x10\x26\x38\x07\xe5\xc7\x36\x80"
"\xe2\x23\x12\x61\xf5\x48\x4b\x2b"
"\xc5\xdf\x15\xd9\x87\x01\xaa\xac"
"\x1e\x7c\xad\x73\x78\x18\x63\xe0"
"\x8b\x9f\x81\xd8\x12\x6a\x28\x10"
"\xbe\x04\x68\x8a\x09\x7c\x1b\x1c"
"\x83\x66\x80\x47\x80\xe8\xfd\x35"
"\x1c\x97\x6f\xae\x49\x10\x66\xcc"
"\xc6\xd8\xcc\x3a\x84\x91\x20\x77"
"\x72\xe4\x24\xd2\x37\x9f\xc5\xc9"
"\x25\x94\x10\x5f\x40\x00\x64\x99"
"\xdc\xae\xd7\x21\x09\x78\x50\x15"
"\xac\x5f\xc6\x2c\xa2\x0b\xa9\x39"
"\x87\x6e\x6d\xab\xde\x08\x51\x16"
"\xc7\x13\xe9\xea\xed\x06\x8e\x2c"
"\xf8\x37\x8c\xf0\xa6\x96\x8d\x43"
"\xb6\x98\x37\xb2\x43\xed\xde\xdf"
"\x89\x1a\xe7\xeb\x9d\xa1\x7b\x0b"
"\x77\xb0\xe2\x75\xc0\xf1\x98\xd9"
"\x80\x55\xc9\x34\x91\xd1\x59\xe8"
"\x4b\x0f\xc1\xa9\x4b\x7a\x84\x06"
"\x20\xa8\x5d\xfa\xd1\xde\x70\x56"
"\x2f\x9e\x91\x9c\x20\xb3\x24\xd8"
"\x84\x3d\xe1\x8c\x7e\x62\x52\xe5"
"\x44\x4b\x9f\xc2\x93\x03\xea\x2b"
"\x59\xc5\xfa\x3f\x91\x2b\xbb\x23"
"\xf5\xb2\x7b\xf5\x38\xaf\xb3\xee"
"\x63\xdc\x7b\xd1\xff\xaa\x8b\xab"
"\x82\x6b\x37\x04\xeb\x74\xbe\x79"
"\xb9\x83\x90\xef\x20\x59\x46\xff"
"\xe9\x97\x3e\x2f\xee\xb6\x64\x18"
"\x38\x4c\x7a\x4a\xf9\x61\xe8\x9a"
"\xa1\xb5\x01\xa6\x47\xd3\x11\xd4"
"\xce\xd3\x91\x49\x88\xc7\xb8\x4d"
"\xb1\xb9\x07\x6d\x16\x72\xae\x46"
"\x5e\x03\xa1\x4b\xb6\x02\x30\xa8"
"\x3d\xa9\x07\x2a\x7c\x19\xe7\x62"
"\x87\xe3\x82\x2f\x6f\xe1\x09\xd9"
"\x94\x97\xea\xdd\x58\x9e\xae\x76"
"\x7e\x35\xe5\xb4\xda\x7e\xf4\xde"
"\xf7\x32\x87\xcd\x93\xbf\x11\x56"
"\x11\xbe\x08\x74\xe1\x69\xad\xe2"
"\xd7\xf8\x86\x75\x8a\x3c\xa4\xbe"
"\x70\xa7\x1b\xfc\x0b\x44\x2a\x76"
"\x35\xea\x5d\x85\x81\xaf\x85\xeb"
"\xa0\x1c\x61\xc2\xf7\x4f\xa5\xdc"
"\x02\x7f\xf6\x95\x40\x6e\x8a\x9a"
"\xf3\x5d\x25\x6e\x14\x3a\x22\xc9"
"\x37\x1c\xeb\x46\x54\x3f\xa5\x91"
"\xc2\xb5\x8c\xfe\x53\x08\x97\x32"
"\x1b\xb2\x30\x27\xfe\x25\x5d\xdc"
"\x08\x87\xd0\xe5\x94\x1a\xd4\xf1"
"\xfe\xd6\xb4\xa3\xe6\x74\x81\x3c"
"\x1b\xb7\x31\xa7\x22\xfd\xd4\xdd"
"\x20\x4e\x7c\x51\xb0\x60\x73\xb8"
"\x9c\xac\x91\x90\x7e\x01\xb0\xe1"
"\x8a\x2f\x75\x1c\x53\x2a\x98\x2a"
"\x06\x52\x95\x52\xb2\xe9\x25\x2e"
"\x4c\xe2\x5a\x00\xb2\x13\x81\x03"
"\x77\x66\x0d\xa5\x99\xda\x4e\x8c"
"\xac\xf3\x13\x53\x27\x45\xaf\x64"
"\x46\xdc\xea\x23\xda\x97\xd1\xab"
"\x7d\x6c\x30\x96\x1f\xbc\x06\x34"
"\x18\x0b\x5e\x21\x35\x11\x8d\x4c"
"\xe0\x2d\xe9\x50\x16\x74\x81\xa8"
"\xb4\x34\xb9\x72\x42\xa6\xcc\xbc"
"\xca\x34\x83\x27\x10\x5b\x68\x45"
"\x8f\x52\x22\x0c\x55\x3d\x29\x7c"
"\xe3\xc0\x66\x05\x42\x91\x5f\x58"
"\xfe\x4a\x62\xd9\x8c\xa9\x04\x19"
"\x04\xa9\x08\x4b\x57\xfc\x67\x53"
"\x08\x7c\xbc\x66\x8a\xb0\xb6\x9f"
"\x92\xd6\x41\x7c\x5b\x2a\x00\x79"
"\x72",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ctext = "\x45\xe8\xe0\xb6\x9c\xca\xfd\x87"
crypto: testmgr - Add a longer ChaCha20 test vector The AVX2 variant of ChaCha20 is used only for messages with >= 512 bytes length. With the existing test vectors, the implementation could not be tested. Due that lack of such a long official test vector, this one is self-generated using chacha20-generic. Signed-off-by: Martin Willi <martin@strongswan.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-16 11:14:04 -06:00
"\xe8\x1d\x37\x96\x8a\xe3\x40\x35"
"\xcf\x5e\x3a\x46\x3d\xfb\xd0\x69"
"\xde\xaf\x7a\xd5\x0d\xe9\x52\xec"
"\xc2\x82\xe5\x3e\x7d\xb2\x4a\xd9"
"\xbb\xc3\x9f\xc0\x5d\xac\x93\x8d"
"\x0e\x6f\xd3\xd7\xfb\x6a\x0d\xce"
"\x92\x2c\xf7\xbb\x93\x57\xcc\xee"
"\x42\x72\x6f\xc8\x4b\xd2\x76\xbf"
"\xa0\xe3\x7a\x39\xf9\x5c\x8e\xfd"
"\xa1\x1d\x41\xe5\x08\xc1\x1c\x11"
"\x92\xfd\x39\x5c\x51\xd0\x2f\x66"
"\x33\x4a\x71\x15\xfe\xee\x12\x54"
"\x8c\x8f\x34\xd8\x50\x3c\x18\xa6"
"\xc5\xe1\x46\x8a\xfb\x5f\x7e\x25"
"\x9b\xe2\xc3\x66\x41\x2b\xb3\xa5"
"\x57\x0e\x94\x17\x26\x39\xbb\x54"
"\xae\x2e\x6f\x42\xfb\x4d\x89\x6f"
"\x9d\xf1\x16\x2e\xe3\xe7\xfc\xe3"
"\xb2\x4b\x2b\xa6\x7c\x04\x69\x3a"
"\x70\x5a\xa7\xf1\x31\x64\x19\xca"
"\x45\x79\xd8\x58\x23\x61\xaf\xc2"
"\x52\x05\xc3\x0b\xc1\x64\x7c\x81"
"\xd9\x11\xcf\xff\x02\x3d\x51\x84"
"\x01\xac\xc6\x2e\x34\x2b\x09\x3a"
"\xa8\x5d\x98\x0e\x89\xd9\xef\x8f"
"\xd9\xd7\x7d\xdd\x63\x47\x46\x7d"
"\xa1\xda\x0b\x53\x7d\x79\xcd\xc9"
"\x86\xdd\x6b\x13\xa1\x9a\x70\xdd"
"\x5c\xa1\x69\x3c\xe4\x5d\xe3\x8c"
"\xe5\xf4\x87\x9c\x10\xcf\x0f\x0b"
"\xc8\x43\xdc\xf8\x1d\x62\x5e\x5b"
"\xe2\x03\x06\xc5\x71\xb6\x48\xa5"
"\xf0\x0f\x2d\xd5\xa2\x73\x55\x8f"
"\x01\xa7\x59\x80\x5f\x11\x6c\x40"
"\xff\xb1\xf2\xc6\x7e\x01\xbb\x1c"
"\x69\x9c\xc9\x3f\x71\x5f\x07\x7e"
"\xdf\x6f\x99\xca\x9c\xfd\xf9\xb9"
"\x49\xe7\xcc\x91\xd5\x9b\x8f\x03"
"\xae\xe7\x61\x32\xef\x41\x6c\x75"
"\x84\x9b\x8c\xce\x1d\x6b\x93\x21"
"\x41\xec\xc6\xad\x8e\x0c\x48\xa8"
"\xe2\xf5\x57\xde\xf7\x38\xfd\x4a"
"\x6f\xa7\x4a\xf9\xac\x7d\xb1\x85"
"\x7d\x6c\x95\x0a\x5a\xcf\x68\xd2"
"\xe0\x7a\x26\xd9\xc1\x6d\x3e\xc6"
"\x37\xbd\xbe\x24\x36\x77\x9f\x1b"
"\xc1\x22\xf3\x79\xae\x95\x78\x66"
"\x97\x11\xc0\x1a\xf1\xe8\x0d\x38"
"\x09\xc2\xee\xb7\xd3\x46\x7b\x59"
"\x77\x23\xe8\xb4\x92\x3d\x78\xbe"
"\xe2\x25\x63\xa5\x2a\x06\x70\x92"
"\x32\x63\xf9\x19\x21\x68\xe1\x0b"
"\x9a\xd0\xee\x21\xdb\x1f\xe0\xde"
"\x3e\x64\x02\x4d\x0e\xe0\x0a\xa9"
"\xed\x19\x8c\xa8\xbf\xe3\x2e\x75"
"\x24\x2b\xb0\xe5\x82\x6a\x1e\x6f"
"\x71\x2a\x3a\x60\xed\x06\x0d\x17"
"\xa2\xdb\x29\x1d\xae\xb2\xc4\xfb"
"\x94\x04\xd8\x58\xfc\xc4\x04\x4e"
"\xee\xc7\xc1\x0f\xe9\x9b\x63\x2d"
"\x02\x3e\x02\x67\xe5\xd8\xbb\x79"
"\xdf\xd2\xeb\x50\xe9\x0a\x02\x46"
"\xdf\x68\xcf\xe7\x2b\x0a\x56\xd6"
"\xf7\xbc\x44\xad\xb8\xb5\x5f\xeb"
"\xbc\x74\x6b\xe8\x7e\xb0\x60\xc6"
"\x0d\x96\x09\xbb\x19\xba\xe0\x3c"
"\xc4\x6c\xbf\x0f\x58\xc0\x55\x62"
"\x23\xa0\xff\xb5\x1c\xfd\x18\xe1"
"\xcf\x6d\xd3\x52\xb4\xce\xa6\xfa"
"\xaa\xfb\x1b\x0b\x42\x6d\x79\x42"
"\x48\x70\x5b\x0e\xdd\x3a\xc9\x69"
"\x8b\x73\x67\xf6\x95\xdb\x8c\xfb"
"\xfd\xb5\x08\x47\x42\x84\x9a\xfa"
"\xcc\x67\xb2\x3c\xb6\xfd\xd8\x32"
"\xd6\x04\xb6\x4a\xea\x53\x4b\xf5"
"\x94\x16\xad\xf0\x10\x2e\x2d\xb4"
"\x8b\xab\xe5\x89\xc7\x39\x12\xf3"
"\x8d\xb5\x96\x0b\x87\x5d\xa7\x7c"
"\xb0\xc2\xf6\x2e\x57\x97\x2c\xdc"
"\x54\x1c\x34\x72\xde\x0c\x68\x39"
"\x9d\x32\xa5\x75\x92\x13\x32\xea"
"\x90\x27\xbd\x5b\x1d\xb9\x21\x02"
"\x1c\xcc\xba\x97\x5e\x49\x58\xe8"
"\xac\x8b\xf3\xce\x3c\xf0\x00\xe9"
"\x6c\xae\xe9\x77\xdf\xf4\x02\xcd"
"\x55\x25\x89\x9e\x90\xf3\x6b\x8f"
"\xb7\xd6\x47\x98\x26\x2f\x31\x2f"
"\x8d\xbf\x54\xcd\x99\xeb\x80\xd7"
"\xac\xc3\x08\xc2\xa6\x32\xf1\x24"
"\x76\x7c\x4f\x78\x53\x55\xfb\x00"
"\x8a\xd6\x52\x53\x25\x45\xfb\x0a"
"\x6b\xb9\xbe\x3c\x5e\x11\xcc\x6a"
"\xdd\xfc\xa7\xc4\x79\x4d\xbd\xfb"
"\xce\x3a\xf1\x7a\xda\xeb\xfe\x64"
"\x28\x3d\x0f\xee\x80\xba\x0c\xf8"
"\xe9\x5b\x3a\xd4\xae\xc9\xf3\x0e"
"\xe8\x5d\xc5\x5c\x0b\x20\x20\xee"
"\x40\x0d\xde\x07\xa7\x14\xb4\x90"
"\xb6\xbd\x3b\xae\x7d\x2b\xa7\xc7"
"\xdc\x0b\x4c\x5d\x65\xb0\xd2\xc5"
"\x79\x61\x23\xe0\xa2\x99\x73\x55"
"\xad\xc6\xfb\xc7\x54\xb5\x98\x1f"
"\x8c\x86\xc2\x3f\xbe\x5e\xea\x64"
"\xa3\x60\x18\x9f\x80\xaf\x52\x74"
"\x1a\xfe\x22\xc2\x92\x67\x40\x02"
"\x08\xee\x67\x5b\x67\xe0\x3d\xde"
"\x7a\xaf\x8e\x28\xf3\x5e\x0e\xf4"
"\x48\x56\xaa\x85\x22\xd8\x36\xed"
"\x3b\x3d\x68\x69\x30\xbc\x71\x23"
"\xb1\x6e\x61\x03\x89\x44\x03\xf4"
"\x32\xaa\x4c\x40\x9f\x69\xfb\x70"
"\x91\xcc\x1f\x11\xbd\x76\x67\xe6"
"\x10\x8b\x29\x39\x68\xea\x4e\x6d"
"\xae\xfb\x40\xcf\xe2\xd0\x0d\x8d"
"\x6f\xed\x9b\x8d\x64\x7a\x94\x8e"
"\x32\x38\x78\xeb\x7d\x5f\xf9\x4d"
"\x13\xbe\x21\xea\x16\xe7\x5c\xee"
"\xcd\xf6\x5f\xc6\x45\xb2\x8f\x2b"
"\xb5\x93\x3e\x45\xdb\xfd\xa2\x6a"
"\xec\x83\x92\x99\x87\x47\xe0\x7c"
"\xa2\x7b\xc4\x2a\xcd\xc0\x81\x03"
"\x98\xb0\x87\xb6\x86\x13\x64\x33"
"\x4c\xd7\x99\xbf\xdb\x7b\x6e\xaa"
"\x76\xcc\xa0\x74\x1b\xa3\x6e\x83"
"\xd4\xba\x7a\x84\x9d\x91\x71\xcd"
"\x60\x2d\x56\xfd\x26\x35\xcb\xeb"
"\xac\xe9\xee\xa4\xfc\x18\x5b\x91"
"\xd5\xfe\x84\x45\xe0\xc7\xfd\x11"
"\xe9\x00\xb6\x54\xdf\xe1\x94\xde"
"\x2b\x70\x9f\x94\x7f\x15\x0e\x83"
"\x63\x10\xb3\xf5\xea\xd3\xe8\xd1"
"\xa5\xfc\x17\x19\x68\x9a\xbc\x17"
"\x30\x43\x0a\x1a\x33\x92\xd4\x2a"
"\x2e\x68\x99\xbc\x49\xf0\x68\xe3"
"\xf0\x1f\xcb\xcc\xfa\xbb\x05\x56"
"\x46\x84\x8b\x69\x83\x64\xc5\xe0"
"\xc5\x52\x99\x07\x3c\xa6\x5c\xaf"
"\xa3\xde\xd7\xdb\x43\xe6\xb7\x76"
"\x4e\x4d\xd6\x71\x60\x63\x4a\x0c"
"\x5f\xae\x25\x84\x22\x90\x5f\x26"
"\x61\x4d\x8f\xaf\xc9\x22\xf2\x05"
"\xcf\xc1\xdc\x68\xe5\x57\x8e\x24"
"\x1b\x30\x59\xca\xd7\x0d\xc3\xd3"
"\x52\x9e\x09\x3e\x0e\xaf\xdb\x5f"
"\xc7\x2b\xde\x3a\xfd\xad\x93\x04"
"\x74\x06\x89\x0e\x90\xeb\x85\xff"
"\xe6\x3c\x12\x42\xf4\xfa\x80\x75"
"\x5e\x4e\xd7\x2f\x93\x0b\x34\x41"
"\x02\x85\x68\xd0\x03\x12\xde\x92"
"\x54\x7a\x7e\xfb\x55\xe7\x88\xfb"
"\xa4\xa9\xf2\xd1\xc6\x70\x06\x37"
"\x25\xee\xa7\x6e\xd9\x89\x86\x50"
"\x2e\x07\xdb\xfb\x2a\x86\x45\x0e"
"\x91\xf4\x7c\xbb\x12\x60\xe8\x3f"
"\x71\xbe\x8f\x9d\x26\xef\xd9\x89"
"\xc4\x8f\xd8\xc5\x73\xd8\x84\xaa"
"\x2f\xad\x22\x1e\x7e\xcf\xa2\x08"
"\x23\x45\x89\x42\xa0\x30\xeb\xbf"
"\xa1\xed\xad\xd5\x76\xfa\x24\x8f"
"\x98",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 1281,
crypto: testmgr - add chunked test cases for chacha20 We failed to catch a bug in the chacha20 code after porting it to the skcipher API. We would have caught it if any chunked tests had been defined, so define some now so we will catch future regressions. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-08-14 07:28:15 -06:00
.also_non_np = 1,
.np = 3,
.tap = { 1200, 1, 80 },
crypto: testmgr - Add ChaCha20 test vectors from RFC7539 We explicitly set the Initial block Counter by prepending it to the nonce in Little Endian. The same test vector is used for both encryption and decryption, ChaCha20 is a cipher XORing a keystream. Signed-off-by: Martin Willi <martin@strongswan.org> Acked-by: Steffen Klassert <steffen.klassert@secunet.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-06-01 05:43:57 -06:00
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* CTS (Cipher Text Stealing) mode tests
*/
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
static const struct cipher_testvec cts_mode_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{ /* from rfc3962 */
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 17,
.ctext = "\xc6\x35\x35\x68\xf2\xbf\x8c\xb4"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xd8\xa5\x80\x36\x2d\xa7\xff\x7f"
"\x97",
}, {
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20\x47\x65\x6e\x65\x72\x61\x6c"
"\x20\x47\x61\x75\x27\x73\x20",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 31,
.ctext = "\xfc\x00\x78\x3e\x0e\xfd\xb2\xc1"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xd4\x45\xd4\xc8\xef\xf7\xed\x22"
"\x97\x68\x72\x68\xd6\xec\xcc\xc0"
"\xc0\x7b\x25\xe2\x5e\xcf\xe5",
}, {
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20\x47\x65\x6e\x65\x72\x61\x6c"
"\x20\x47\x61\x75\x27\x73\x20\x43",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 32,
.ctext = "\x39\x31\x25\x23\xa7\x86\x62\xd5"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xbe\x7f\xcb\xcc\x98\xeb\xf5\xa8"
"\x97\x68\x72\x68\xd6\xec\xcc\xc0"
"\xc0\x7b\x25\xe2\x5e\xcf\xe5\x84",
}, {
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20\x47\x65\x6e\x65\x72\x61\x6c"
"\x20\x47\x61\x75\x27\x73\x20\x43"
"\x68\x69\x63\x6b\x65\x6e\x2c\x20"
"\x70\x6c\x65\x61\x73\x65\x2c",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 47,
.ctext = "\x97\x68\x72\x68\xd6\xec\xcc\xc0"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xc0\x7b\x25\xe2\x5e\xcf\xe5\x84"
"\xb3\xff\xfd\x94\x0c\x16\xa1\x8c"
"\x1b\x55\x49\xd2\xf8\x38\x02\x9e"
"\x39\x31\x25\x23\xa7\x86\x62\xd5"
"\xbe\x7f\xcb\xcc\x98\xeb\xf5",
}, {
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20\x47\x65\x6e\x65\x72\x61\x6c"
"\x20\x47\x61\x75\x27\x73\x20\x43"
"\x68\x69\x63\x6b\x65\x6e\x2c\x20"
"\x70\x6c\x65\x61\x73\x65\x2c\x20",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 48,
.ctext = "\x97\x68\x72\x68\xd6\xec\xcc\xc0"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xc0\x7b\x25\xe2\x5e\xcf\xe5\x84"
"\x9d\xad\x8b\xbb\x96\xc4\xcd\xc0"
"\x3b\xc1\x03\xe1\xa1\x94\xbb\xd8"
"\x39\x31\x25\x23\xa7\x86\x62\xd5"
"\xbe\x7f\xcb\xcc\x98\xeb\xf5\xa8",
}, {
.klen = 16,
.key = "\x63\x68\x69\x63\x6b\x65\x6e\x20"
"\x74\x65\x72\x69\x79\x61\x6b\x69",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.ptext = "\x49\x20\x77\x6f\x75\x6c\x64\x20"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6c\x69\x6b\x65\x20\x74\x68\x65"
"\x20\x47\x65\x6e\x65\x72\x61\x6c"
"\x20\x47\x61\x75\x27\x73\x20\x43"
"\x68\x69\x63\x6b\x65\x6e\x2c\x20"
"\x70\x6c\x65\x61\x73\x65\x2c\x20"
"\x61\x6e\x64\x20\x77\x6f\x6e\x74"
"\x6f\x6e\x20\x73\x6f\x75\x70\x2e",
crypto: testmgr - eliminate redundant decryption test vectors Currently testmgr has separate encryption and decryption test vectors for symmetric ciphers. That's massively redundant, since with few exceptions (mostly mistakes, apparently), all decryption tests are identical to the encryption tests, just with the input/result flipped. Therefore, eliminate the redundancy by removing the decryption test vectors and updating testmgr to test both encryption and decryption using what used to be the encryption test vectors. Naming is adjusted accordingly: each cipher_testvec now has a 'ptext' (plaintext), 'ctext' (ciphertext), and 'len' instead of an 'input', 'result', 'ilen', and 'rlen'. Note that it was always the case that 'ilen == rlen'. AES keywrap ("kw(aes)") is special because its IV is generated by the encryption. Previously this was handled by specifying 'iv_out' for encryption and 'iv' for decryption. To make it work cleanly with only one set of test vectors, put the IV in 'iv', remove 'iv_out', and add a boolean that indicates that the IV is generated by the encryption. In total, this removes over 10000 lines from testmgr.h, with no reduction in test coverage since prior patches already copied the few unique decryption test vectors into the encryption test vectors. This covers all algorithms that used 'struct cipher_testvec', e.g. any block cipher in the ECB, CBC, CTR, XTS, LRW, CTS-CBC, PCBC, OFB, or keywrap modes, and Salsa20 and ChaCha20. No change is made to AEAD tests, though we probably can eliminate a similar redundancy there too. The testmgr.h portion of this patch was automatically generated using the following awk script, with some slight manual fixups on top (updated 'struct cipher_testvec' definition, updated a few comments, and fixed up the AES keywrap test vectors): BEGIN { OTHER = 0; ENCVEC = 1; DECVEC = 2; DECVEC_TAIL = 3; mode = OTHER } /^static const struct cipher_testvec.*_enc_/ { sub("_enc", ""); mode = ENCVEC } /^static const struct cipher_testvec.*_dec_/ { mode = DECVEC } mode == ENCVEC && !/\.ilen[[:space:]]*=/ { sub(/\.input[[:space:]]*=$/, ".ptext =") sub(/\.input[[:space:]]*=/, ".ptext\t=") sub(/\.result[[:space:]]*=$/, ".ctext =") sub(/\.result[[:space:]]*=/, ".ctext\t=") sub(/\.rlen[[:space:]]*=/, ".len\t=") print } mode == DECVEC_TAIL && /[^[:space:]]/ { mode = OTHER } mode == OTHER { print } mode == ENCVEC && /^};/ { mode = OTHER } mode == DECVEC && /^};/ { mode = DECVEC_TAIL } Note that git's default diff algorithm gets confused by the testmgr.h portion of this patch, and reports too many lines added and removed. It's better viewed with 'git diff --minimal' (or 'git show --minimal'), which reports "2 files changed, 919 insertions(+), 11723 deletions(-)". Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-20 23:50:29 -06:00
.len = 64,
.ctext = "\x97\x68\x72\x68\xd6\xec\xcc\xc0"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\xc0\x7b\x25\xe2\x5e\xcf\xe5\x84"
"\x39\x31\x25\x23\xa7\x86\x62\xd5"
"\xbe\x7f\xcb\xcc\x98\xeb\xf5\xa8"
"\x48\x07\xef\xe8\x36\xee\x89\xa5"
"\x26\x73\x0d\xbc\x2f\x7b\xc8\x40"
"\x9d\xad\x8b\xbb\x96\xc4\xcd\xc0"
"\x3b\xc1\x03\xe1\xa1\x94\xbb\xd8",
}
};
/*
* Compression stuff.
*/
#define COMP_BUF_SIZE 512
struct comp_testvec {
int inlen, outlen;
char input[COMP_BUF_SIZE];
char output[COMP_BUF_SIZE];
};
/*
* Deflate test vectors (null-terminated strings).
crypto: testmgr - Correct comment about deflate parameters The comment for the deflate test vectors says the winbits parameter is 11, while the deflate module actually uses -11 (a negative window bits parameter enables the raw deflate format instead of the zlib format). Correct this, to avoid confusion about the format used. Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-12-17 23:17:46 -07:00
* Params: winbits=-11, Z_DEFAULT_COMPRESSION, MAX_MEM_LEVEL.
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
*/
crypto: testmgr - add zlib test Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-03-04 00:42:15 -07:00
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec deflate_comp_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.inlen = 70,
.outlen = 38,
.input = "Join us now and share the software "
"Join us now and share the software ",
.output = "\xf3\xca\xcf\xcc\x53\x28\x2d\x56"
"\xc8\xcb\x2f\x57\x48\xcc\x4b\x51"
"\x28\xce\x48\x2c\x4a\x55\x28\xc9"
"\x48\x55\x28\xce\x4f\x2b\x29\x07"
"\x71\xbc\x08\x2b\x01\x00",
}, {
.inlen = 191,
.outlen = 122,
.input = "This document describes a compression method based on the DEFLATE"
"compression algorithm. This document defines the application of "
"the DEFLATE algorithm to the IP Payload Compression Protocol.",
.output = "\x5d\x8d\x31\x0e\xc2\x30\x10\x04"
"\xbf\xb2\x2f\xc8\x1f\x10\x04\x09"
"\x89\xc2\x85\x3f\x70\xb1\x2f\xf8"
"\x24\xdb\x67\xd9\x47\xc1\xef\x49"
"\x68\x12\x51\xae\x76\x67\xd6\x27"
"\x19\x88\x1a\xde\x85\xab\x21\xf2"
"\x08\x5d\x16\x1e\x20\x04\x2d\xad"
"\xf3\x18\xa2\x15\x85\x2d\x69\xc4"
"\x42\x83\x23\xb6\x6c\x89\x71\x9b"
"\xef\xcf\x8b\x9f\xcf\x33\xca\x2f"
"\xed\x62\xa9\x4c\x80\xff\x13\xaf"
"\x52\x37\xed\x0e\x52\x6b\x59\x02"
"\xd9\x4e\xe8\x7a\x76\x1d\x02\x98"
"\xfe\x8a\x87\x83\xa3\x4f\x56\x8a"
"\xb8\x9e\x8e\x5c\x57\xd3\xa0\x79"
"\xfa\x02",
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec deflate_decomp_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.inlen = 122,
.outlen = 191,
.input = "\x5d\x8d\x31\x0e\xc2\x30\x10\x04"
"\xbf\xb2\x2f\xc8\x1f\x10\x04\x09"
"\x89\xc2\x85\x3f\x70\xb1\x2f\xf8"
"\x24\xdb\x67\xd9\x47\xc1\xef\x49"
"\x68\x12\x51\xae\x76\x67\xd6\x27"
"\x19\x88\x1a\xde\x85\xab\x21\xf2"
"\x08\x5d\x16\x1e\x20\x04\x2d\xad"
"\xf3\x18\xa2\x15\x85\x2d\x69\xc4"
"\x42\x83\x23\xb6\x6c\x89\x71\x9b"
"\xef\xcf\x8b\x9f\xcf\x33\xca\x2f"
"\xed\x62\xa9\x4c\x80\xff\x13\xaf"
"\x52\x37\xed\x0e\x52\x6b\x59\x02"
"\xd9\x4e\xe8\x7a\x76\x1d\x02\x98"
"\xfe\x8a\x87\x83\xa3\x4f\x56\x8a"
"\xb8\x9e\x8e\x5c\x57\xd3\xa0\x79"
"\xfa\x02",
.output = "This document describes a compression method based on the DEFLATE"
"compression algorithm. This document defines the application of "
"the DEFLATE algorithm to the IP Payload Compression Protocol.",
}, {
.inlen = 38,
.outlen = 70,
.input = "\xf3\xca\xcf\xcc\x53\x28\x2d\x56"
"\xc8\xcb\x2f\x57\x48\xcc\x4b\x51"
"\x28\xce\x48\x2c\x4a\x55\x28\xc9"
"\x48\x55\x28\xce\x4f\x2b\x29\x07"
crypto: testmgr - add zlib test Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2009-03-04 00:42:15 -07:00
"\x71\xbc\x08\x2b\x01\x00",
.output = "Join us now and share the software "
"Join us now and share the software ",
},
};
crypto: scomp - add support for deflate rfc1950 (zlib) Add scomp backend for zlib-deflate compression algorithm. This backend outputs data using the format defined in rfc1950 (raw deflate surrounded by zlib header and footer). Signed-off-by: Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-04-21 14:54:30 -06:00
static const struct comp_testvec zlib_deflate_comp_tv_template[] = {
{
.inlen = 70,
.outlen = 44,
.input = "Join us now and share the software "
"Join us now and share the software ",
.output = "\x78\x5e\xf3\xca\xcf\xcc\x53\x28"
"\x2d\x56\xc8\xcb\x2f\x57\x48\xcc"
"\x4b\x51\x28\xce\x48\x2c\x4a\x55"
"\x28\xc9\x48\x55\x28\xce\x4f\x2b"
"\x29\x07\x71\xbc\x08\x2b\x01\x00"
"\x7c\x65\x19\x3d",
}, {
.inlen = 191,
.outlen = 129,
.input = "This document describes a compression method based on the DEFLATE"
"compression algorithm. This document defines the application of "
"the DEFLATE algorithm to the IP Payload Compression Protocol.",
.output = "\x78\x5e\x5d\xce\x41\x0a\xc3\x30"
"\x0c\x04\xc0\xaf\xec\x0b\xf2\x87"
"\xd2\xa6\x50\xe8\xc1\x07\x7f\x40"
"\xb1\x95\x5a\x60\x5b\xc6\x56\x0f"
"\xfd\x7d\x93\x1e\x42\xe8\x51\xec"
"\xee\x20\x9f\x64\x20\x6a\x78\x17"
"\xae\x86\xc8\x23\x74\x59\x78\x80"
"\x10\xb4\xb4\xce\x63\x88\x56\x14"
"\xb6\xa4\x11\x0b\x0d\x8e\xd8\x6e"
"\x4b\x8c\xdb\x7c\x7f\x5e\xfc\x7c"
"\xae\x51\x7e\x69\x17\x4b\x65\x02"
"\xfc\x1f\xbc\x4a\xdd\xd8\x7d\x48"
"\xad\x65\x09\x64\x3b\xac\xeb\xd9"
"\xc2\x01\xc0\xf4\x17\x3c\x1c\x1c"
"\x7d\xb2\x52\xc4\xf5\xf4\x8f\xeb"
"\x6a\x1a\x34\x4f\x5f\x2e\x32\x45"
"\x4e",
},
};
static const struct comp_testvec zlib_deflate_decomp_tv_template[] = {
{
.inlen = 128,
.outlen = 191,
.input = "\x78\x9c\x5d\x8d\x31\x0e\xc2\x30"
"\x10\x04\xbf\xb2\x2f\xc8\x1f\x10"
"\x04\x09\x89\xc2\x85\x3f\x70\xb1"
"\x2f\xf8\x24\xdb\x67\xd9\x47\xc1"
"\xef\x49\x68\x12\x51\xae\x76\x67"
"\xd6\x27\x19\x88\x1a\xde\x85\xab"
"\x21\xf2\x08\x5d\x16\x1e\x20\x04"
"\x2d\xad\xf3\x18\xa2\x15\x85\x2d"
"\x69\xc4\x42\x83\x23\xb6\x6c\x89"
"\x71\x9b\xef\xcf\x8b\x9f\xcf\x33"
"\xca\x2f\xed\x62\xa9\x4c\x80\xff"
"\x13\xaf\x52\x37\xed\x0e\x52\x6b"
"\x59\x02\xd9\x4e\xe8\x7a\x76\x1d"
"\x02\x98\xfe\x8a\x87\x83\xa3\x4f"
"\x56\x8a\xb8\x9e\x8e\x5c\x57\xd3"
"\xa0\x79\xfa\x02\x2e\x32\x45\x4e",
.output = "This document describes a compression method based on the DEFLATE"
"compression algorithm. This document defines the application of "
"the DEFLATE algorithm to the IP Payload Compression Protocol.",
}, {
.inlen = 44,
.outlen = 70,
.input = "\x78\x9c\xf3\xca\xcf\xcc\x53\x28"
"\x2d\x56\xc8\xcb\x2f\x57\x48\xcc"
"\x4b\x51\x28\xce\x48\x2c\x4a\x55"
"\x28\xc9\x48\x55\x28\xce\x4f\x2b"
"\x29\x07\x71\xbc\x08\x2b\x01\x00"
"\x7c\x65\x19\x3d",
.output = "Join us now and share the software "
"Join us now and share the software ",
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* LZO test vectors (null-terminated strings).
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lzo_comp_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.inlen = 70,
crypto: testmgr - update LZO compression test vectors Update the LZO compression test vectors according to the latest compressor version. Signed-off-by: Markus F.X.J. Oberhumer <markus@oberhumer.com>
2012-10-14 07:39:04 -06:00
.outlen = 57,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.input = "Join us now and share the software "
"Join us now and share the software ",
.output = "\x00\x0d\x4a\x6f\x69\x6e\x20\x75"
crypto: testmgr - update LZO compression test vectors Update the LZO compression test vectors according to the latest compressor version. Signed-off-by: Markus F.X.J. Oberhumer <markus@oberhumer.com>
2012-10-14 07:39:04 -06:00
"\x73\x20\x6e\x6f\x77\x20\x61\x6e"
"\x64\x20\x73\x68\x61\x72\x65\x20"
"\x74\x68\x65\x20\x73\x6f\x66\x74"
"\x77\x70\x01\x32\x88\x00\x0c\x65"
"\x20\x74\x68\x65\x20\x73\x6f\x66"
"\x74\x77\x61\x72\x65\x20\x11\x00"
"\x00",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
}, {
.inlen = 159,
crypto: testmgr - update LZO compression test vectors Update the LZO compression test vectors according to the latest compressor version. Signed-off-by: Markus F.X.J. Oberhumer <markus@oberhumer.com>
2012-10-14 07:39:04 -06:00
.outlen = 131,
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
.input = "This document describes a compression method based on the LZO "
"compression algorithm. This document defines the application of "
"the LZO algorithm used in UBIFS.",
crypto: testmgr - update LZO compression test vectors Update the LZO compression test vectors according to the latest compressor version. Signed-off-by: Markus F.X.J. Oberhumer <markus@oberhumer.com>
2012-10-14 07:39:04 -06:00
.output = "\x00\x2c\x54\x68\x69\x73\x20\x64"
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
"\x6f\x63\x75\x6d\x65\x6e\x74\x20"
"\x64\x65\x73\x63\x72\x69\x62\x65"
"\x73\x20\x61\x20\x63\x6f\x6d\x70"
"\x72\x65\x73\x73\x69\x6f\x6e\x20"
"\x6d\x65\x74\x68\x6f\x64\x20\x62"
"\x61\x73\x65\x64\x20\x6f\x6e\x20"
crypto: testmgr - update LZO compression test vectors Update the LZO compression test vectors according to the latest compressor version. Signed-off-by: Markus F.X.J. Oberhumer <markus@oberhumer.com>
2012-10-14 07:39:04 -06:00
"\x74\x68\x65\x20\x4c\x5a\x4f\x20"
"\x2a\x8c\x00\x09\x61\x6c\x67\x6f"
"\x72\x69\x74\x68\x6d\x2e\x20\x20"
"\x2e\x54\x01\x03\x66\x69\x6e\x65"
"\x73\x20\x74\x06\x05\x61\x70\x70"
"\x6c\x69\x63\x61\x74\x76\x0a\x6f"
"\x66\x88\x02\x60\x09\x27\xf0\x00"
"\x0c\x20\x75\x73\x65\x64\x20\x69"
"\x6e\x20\x55\x42\x49\x46\x53\x2e"
"\x11\x00\x00",
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lzo_decomp_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.inlen = 133,
.outlen = 159,
.input = "\x00\x2b\x54\x68\x69\x73\x20\x64"
"\x6f\x63\x75\x6d\x65\x6e\x74\x20"
"\x64\x65\x73\x63\x72\x69\x62\x65"
"\x73\x20\x61\x20\x63\x6f\x6d\x70"
"\x72\x65\x73\x73\x69\x6f\x6e\x20"
"\x6d\x65\x74\x68\x6f\x64\x20\x62"
"\x61\x73\x65\x64\x20\x6f\x6e\x20"
"\x74\x68\x65\x20\x4c\x5a\x4f\x2b"
"\x8c\x00\x0d\x61\x6c\x67\x6f\x72"
"\x69\x74\x68\x6d\x2e\x20\x20\x54"
"\x68\x69\x73\x2a\x54\x01\x02\x66"
"\x69\x6e\x65\x73\x94\x06\x05\x61"
"\x70\x70\x6c\x69\x63\x61\x74\x76"
"\x0a\x6f\x66\x88\x02\x60\x09\x27"
"\xf0\x00\x0c\x20\x75\x73\x65\x64"
"\x20\x69\x6e\x20\x55\x42\x49\x46"
"\x53\x2e\x11\x00\x00",
.output = "This document describes a compression method based on the LZO "
"compression algorithm. This document defines the application of "
"the LZO algorithm used in UBIFS.",
}, {
.inlen = 46,
.outlen = 70,
.input = "\x00\x0d\x4a\x6f\x69\x6e\x20\x75"
"\x73\x20\x6e\x6f\x77\x20\x61\x6e"
"\x64\x20\x73\x68\x61\x72\x65\x20"
"\x74\x68\x65\x20\x73\x6f\x66\x74"
"\x77\x70\x01\x01\x4a\x6f\x69\x6e"
"\x3d\x88\x00\x11\x00\x00",
.output = "Join us now and share the software "
"Join us now and share the software ",
},
};
/*
* Michael MIC test vectors from IEEE 802.11i
*/
#define MICHAEL_MIC_TEST_VECTORS 6
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec michael_mic_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x00\x00\x00\x00\x00\x00\x00\x00",
.ksize = 8,
.plaintext = zeroed_string,
.psize = 0,
.digest = "\x82\x92\x5c\x1c\xa1\xd1\x30\xb8",
},
{
.key = "\x82\x92\x5c\x1c\xa1\xd1\x30\xb8",
.ksize = 8,
.plaintext = "M",
.psize = 1,
.digest = "\x43\x47\x21\xca\x40\x63\x9b\x3f",
},
{
.key = "\x43\x47\x21\xca\x40\x63\x9b\x3f",
.ksize = 8,
.plaintext = "Mi",
.psize = 2,
.digest = "\xe8\xf9\xbe\xca\xe9\x7e\x5d\x29",
},
{
.key = "\xe8\xf9\xbe\xca\xe9\x7e\x5d\x29",
.ksize = 8,
.plaintext = "Mic",
.psize = 3,
.digest = "\x90\x03\x8f\xc6\xcf\x13\xc1\xdb",
},
{
.key = "\x90\x03\x8f\xc6\xcf\x13\xc1\xdb",
.ksize = 8,
.plaintext = "Mich",
.psize = 4,
.digest = "\xd5\x5e\x10\x05\x10\x12\x89\x86",
},
{
.key = "\xd5\x5e\x10\x05\x10\x12\x89\x86",
.ksize = 8,
.plaintext = "Michael",
.psize = 7,
.digest = "\x0a\x94\x2b\x12\x4e\xca\xa5\x46",
}
};
crypto: testmgr - add test cases for CRC32 This adds a couple of test cases for CRC32 (not CRC32c) to ensure that the generic and arch specific implementations are in sync. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Acked-by: Steve Capper <steve.capper@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-05-04 03:00:17 -06:00
/*
* CRC32 test vectors
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec crc32_tv_template[] = {
crypto: testmgr - add more unkeyed crc32 and crc32c test vectors crc32c has an unkeyed test vector but crc32 did not. Add the crc32c one (which uses an empty input) to crc32 too, and also add a new one to both that uses a nonempty input. These test vectors verify that crc32 and crc32c implementations use the correct default initial state. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-19 23:07:42 -06:00
{
.psize = 0,
.digest = "\x00\x00\x00\x00",
},
{
.plaintext = "abcdefg",
.psize = 7,
.digest = "\xd8\xb5\x46\xac",
},
crypto: testmgr - add test cases for CRC32 This adds a couple of test cases for CRC32 (not CRC32c) to ensure that the generic and arch specific implementations are in sync. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> Acked-by: Steve Capper <steve.capper@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-05-04 03:00:17 -06:00
{
.key = "\x87\xa9\xcb\xed",
.ksize = 4,
.psize = 0,
.digest = "\x87\xa9\xcb\xed",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
"\x21\x22\x23\x24\x25\x26\x27\x28",
.psize = 40,
.digest = "\x3a\xdf\x4b\xb0",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
.psize = 40,
.digest = "\xa9\x7a\x7f\x7b",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78",
.psize = 40,
.digest = "\xba\xd3\xf8\x1c",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
.psize = 40,
.digest = "\xa8\xa9\xc2\x02",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
.psize = 40,
.digest = "\x27\xf0\x57\xe2",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 40,
.digest = "\x49\x78\x10\x08",
},
{
.key = "\x80\xea\xd3\xf1",
.ksize = 4,
.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
.psize = 40,
.digest = "\x9a\xb1\xdc\xf0",
},
{
.key = "\xf3\x4a\x1d\x5d",
.ksize = 4,
.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78",
.psize = 40,
.digest = "\xb4\x97\xcc\xd4",
},
{
.key = "\x2e\x80\x04\x59",
.ksize = 4,
.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
.psize = 40,
.digest = "\x67\x9b\xfa\x79",
},
{
.key = "\xa6\xcc\x19\x85",
.ksize = 4,
.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
.psize = 40,
.digest = "\x24\xb5\x16\xef",
},
{
.key = "\x41\xfc\xfe\x2d",
.ksize = 4,
.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 40,
.digest = "\x15\x94\x80\x39",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
"\x21\x22\x23\x24\x25\x26\x27\x28"
"\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50"
"\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78"
"\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0"
"\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8"
"\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 240,
.digest = "\x6c\xc6\x56\xde",
.np = 2,
.tap = { 31, 209 }
}, {
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x6e\x05\x79\x10\xa7\x1b\xb2\x49"
"\xe0\x54\xeb\x82\x19\x8d\x24\xbb"
"\x2f\xc6\x5d\xf4\x68\xff\x96\x0a"
"\xa1\x38\xcf\x43\xda\x71\x08\x7c"
"\x13\xaa\x1e\xb5\x4c\xe3\x57\xee"
"\x85\x1c\x90\x27\xbe\x32\xc9\x60"
"\xf7\x6b\x02\x99\x0d\xa4\x3b\xd2"
"\x46\xdd\x74\x0b\x7f\x16\xad\x21"
"\xb8\x4f\xe6\x5a\xf1\x88\x1f\x93"
"\x2a\xc1\x35\xcc\x63\xfa\x6e\x05"
"\x9c\x10\xa7\x3e\xd5\x49\xe0\x77"
"\x0e\x82\x19\xb0\x24\xbb\x52\xe9"
"\x5d\xf4\x8b\x22\x96\x2d\xc4\x38"
"\xcf\x66\xfd\x71\x08\x9f\x13\xaa"
"\x41\xd8\x4c\xe3\x7a\x11\x85\x1c"
"\xb3\x27\xbe\x55\xec\x60\xf7\x8e"
"\x02\x99\x30\xc7\x3b\xd2\x69\x00"
"\x74\x0b\xa2\x16\xad\x44\xdb\x4f"
"\xe6\x7d\x14\x88\x1f\xb6\x2a\xc1"
"\x58\xef\x63\xfa\x91\x05\x9c\x33"
"\xca\x3e\xd5\x6c\x03\x77\x0e\xa5"
"\x19\xb0\x47\xde\x52\xe9\x80\x17"
"\x8b\x22\xb9\x2d\xc4\x5b\xf2\x66"
"\xfd\x94\x08\x9f\x36\xcd\x41\xd8"
"\x6f\x06\x7a\x11\xa8\x1c\xb3\x4a"
"\xe1\x55\xec\x83\x1a\x8e\x25\xbc"
"\x30\xc7\x5e\xf5\x69\x00\x97\x0b"
"\xa2\x39\xd0\x44\xdb\x72\x09\x7d"
"\x14\xab\x1f\xb6\x4d\xe4\x58\xef"
"\x86\x1d\x91\x28\xbf\x33\xca\x61"
"\xf8\x6c\x03\x9a\x0e\xa5\x3c\xd3"
"\x47\xde\x75\x0c\x80\x17\xae\x22"
"\xb9\x50\xe7\x5b\xf2\x89\x20\x94"
"\x2b\xc2\x36\xcd\x64\xfb\x6f\x06"
"\x9d\x11\xa8\x3f\xd6\x4a\xe1\x78"
"\x0f\x83\x1a\xb1\x25\xbc\x53\xea"
"\x5e\xf5\x8c\x00\x97\x2e\xc5\x39"
"\xd0\x67\xfe\x72\x09\xa0\x14\xab"
"\x42\xd9\x4d\xe4\x7b\x12\x86\x1d"
"\xb4\x28\xbf\x56\xed\x61\xf8\x8f"
"\x03\x9a\x31\xc8\x3c\xd3\x6a\x01"
"\x75\x0c\xa3\x17\xae\x45\xdc\x50"
"\xe7\x7e\x15\x89\x20\xb7\x2b\xc2"
"\x59\xf0\x64\xfb\x92\x06\x9d\x34"
"\xcb\x3f\xd6\x6d\x04\x78\x0f\xa6"
"\x1a\xb1\x48\xdf\x53\xea\x81\x18"
"\x8c\x23\xba\x2e\xc5\x5c\xf3\x67"
"\xfe\x95\x09\xa0\x37\xce\x42\xd9"
"\x70\x07\x7b\x12\xa9\x1d\xb4\x4b"
"\xe2\x56\xed\x84\x1b\x8f\x26\xbd"
"\x31\xc8\x5f\xf6\x6a\x01\x98\x0c"
"\xa3\x3a\xd1\x45\xdc\x73\x0a\x7e"
"\x15\xac\x20\xb7\x4e\xe5\x59\xf0"
"\x87\x1e\x92\x29\xc0\x34\xcb\x62"
"\xf9\x6d\x04\x9b\x0f\xa6\x3d\xd4"
"\x48\xdf\x76\x0d\x81\x18\xaf\x23"
"\xba\x51\xe8\x5c\xf3\x8a\x21\x95"
"\x2c\xc3\x37\xce\x65\xfc\x70\x07"
"\x9e\x12\xa9\x40\xd7\x4b\xe2\x79"
"\x10\x84\x1b\xb2\x26\xbd\x54\xeb"
"\x5f\xf6\x8d\x01\x98\x2f\xc6\x3a"
"\xd1\x68\xff\x73\x0a\xa1\x15\xac"
"\x43\xda\x4e\xe5\x7c\x13\x87\x1e"
"\xb5\x29\xc0\x57\xee\x62\xf9\x90"
"\x04\x9b\x32\xc9\x3d\xd4\x6b\x02"
"\x76\x0d\xa4\x18\xaf\x46\xdd\x51"
"\xe8\x7f\x16\x8a\x21\xb8\x2c\xc3"
"\x5a\xf1\x65\xfc\x93\x07\x9e\x35"
"\xcc\x40\xd7\x6e\x05\x79\x10\xa7"
"\x1b\xb2\x49\xe0\x54\xeb\x82\x19"
"\x8d\x24\xbb\x2f\xc6\x5d\xf4\x68"
"\xff\x96\x0a\xa1\x38\xcf\x43\xda"
"\x71\x08\x7c\x13\xaa\x1e\xb5\x4c"
"\xe3\x57\xee\x85\x1c\x90\x27\xbe"
"\x32\xc9\x60\xf7\x6b\x02\x99\x0d"
"\xa4\x3b\xd2\x46\xdd\x74\x0b\x7f"
"\x16\xad\x21\xb8\x4f\xe6\x5a\xf1"
"\x88\x1f\x93\x2a\xc1\x35\xcc\x63"
"\xfa\x6e\x05\x9c\x10\xa7\x3e\xd5"
"\x49\xe0\x77\x0e\x82\x19\xb0\x24"
"\xbb\x52\xe9\x5d\xf4\x8b\x22\x96"
"\x2d\xc4\x38\xcf\x66\xfd\x71\x08"
"\x9f\x13\xaa\x41\xd8\x4c\xe3\x7a"
"\x11\x85\x1c\xb3\x27\xbe\x55\xec"
"\x60\xf7\x8e\x02\x99\x30\xc7\x3b"
"\xd2\x69\x00\x74\x0b\xa2\x16\xad"
"\x44\xdb\x4f\xe6\x7d\x14\x88\x1f"
"\xb6\x2a\xc1\x58\xef\x63\xfa\x91"
"\x05\x9c\x33\xca\x3e\xd5\x6c\x03"
"\x77\x0e\xa5\x19\xb0\x47\xde\x52"
"\xe9\x80\x17\x8b\x22\xb9\x2d\xc4"
"\x5b\xf2\x66\xfd\x94\x08\x9f\x36"
"\xcd\x41\xd8\x6f\x06\x7a\x11\xa8"
"\x1c\xb3\x4a\xe1\x55\xec\x83\x1a"
"\x8e\x25\xbc\x30\xc7\x5e\xf5\x69"
"\x00\x97\x0b\xa2\x39\xd0\x44\xdb"
"\x72\x09\x7d\x14\xab\x1f\xb6\x4d"
"\xe4\x58\xef\x86\x1d\x91\x28\xbf"
"\x33\xca\x61\xf8\x6c\x03\x9a\x0e"
"\xa5\x3c\xd3\x47\xde\x75\x0c\x80"
"\x17\xae\x22\xb9\x50\xe7\x5b\xf2"
"\x89\x20\x94\x2b\xc2\x36\xcd\x64"
"\xfb\x6f\x06\x9d\x11\xa8\x3f\xd6"
"\x4a\xe1\x78\x0f\x83\x1a\xb1\x25"
"\xbc\x53\xea\x5e\xf5\x8c\x00\x97"
"\x2e\xc5\x39\xd0\x67\xfe\x72\x09"
"\xa0\x14\xab\x42\xd9\x4d\xe4\x7b"
"\x12\x86\x1d\xb4\x28\xbf\x56\xed"
"\x61\xf8\x8f\x03\x9a\x31\xc8\x3c"
"\xd3\x6a\x01\x75\x0c\xa3\x17\xae"
"\x45\xdc\x50\xe7\x7e\x15\x89\x20"
"\xb7\x2b\xc2\x59\xf0\x64\xfb\x92"
"\x06\x9d\x34\xcb\x3f\xd6\x6d\x04"
"\x78\x0f\xa6\x1a\xb1\x48\xdf\x53"
"\xea\x81\x18\x8c\x23\xba\x2e\xc5"
"\x5c\xf3\x67\xfe\x95\x09\xa0\x37"
"\xce\x42\xd9\x70\x07\x7b\x12\xa9"
"\x1d\xb4\x4b\xe2\x56\xed\x84\x1b"
"\x8f\x26\xbd\x31\xc8\x5f\xf6\x6a"
"\x01\x98\x0c\xa3\x3a\xd1\x45\xdc"
"\x73\x0a\x7e\x15\xac\x20\xb7\x4e"
"\xe5\x59\xf0\x87\x1e\x92\x29\xc0"
"\x34\xcb\x62\xf9\x6d\x04\x9b\x0f"
"\xa6\x3d\xd4\x48\xdf\x76\x0d\x81"
"\x18\xaf\x23\xba\x51\xe8\x5c\xf3"
"\x8a\x21\x95\x2c\xc3\x37\xce\x65"
"\xfc\x70\x07\x9e\x12\xa9\x40\xd7"
"\x4b\xe2\x79\x10\x84\x1b\xb2\x26"
"\xbd\x54\xeb\x5f\xf6\x8d\x01\x98"
"\x2f\xc6\x3a\xd1\x68\xff\x73\x0a"
"\xa1\x15\xac\x43\xda\x4e\xe5\x7c"
"\x13\x87\x1e\xb5\x29\xc0\x57\xee"
"\x62\xf9\x90\x04\x9b\x32\xc9\x3d"
"\xd4\x6b\x02\x76\x0d\xa4\x18\xaf"
"\x46\xdd\x51\xe8\x7f\x16\x8a\x21"
"\xb8\x2c\xc3\x5a\xf1\x65\xfc\x93"
"\x07\x9e\x35\xcc\x40\xd7\x6e\x05"
"\x79\x10\xa7\x1b\xb2\x49\xe0\x54"
"\xeb\x82\x19\x8d\x24\xbb\x2f\xc6"
"\x5d\xf4\x68\xff\x96\x0a\xa1\x38"
"\xcf\x43\xda\x71\x08\x7c\x13\xaa"
"\x1e\xb5\x4c\xe3\x57\xee\x85\x1c"
"\x90\x27\xbe\x32\xc9\x60\xf7\x6b"
"\x02\x99\x0d\xa4\x3b\xd2\x46\xdd"
"\x74\x0b\x7f\x16\xad\x21\xb8\x4f"
"\xe6\x5a\xf1\x88\x1f\x93\x2a\xc1"
"\x35\xcc\x63\xfa\x6e\x05\x9c\x10"
"\xa7\x3e\xd5\x49\xe0\x77\x0e\x82"
"\x19\xb0\x24\xbb\x52\xe9\x5d\xf4"
"\x8b\x22\x96\x2d\xc4\x38\xcf\x66"
"\xfd\x71\x08\x9f\x13\xaa\x41\xd8"
"\x4c\xe3\x7a\x11\x85\x1c\xb3\x27"
"\xbe\x55\xec\x60\xf7\x8e\x02\x99"
"\x30\xc7\x3b\xd2\x69\x00\x74\x0b"
"\xa2\x16\xad\x44\xdb\x4f\xe6\x7d"
"\x14\x88\x1f\xb6\x2a\xc1\x58\xef"
"\x63\xfa\x91\x05\x9c\x33\xca\x3e"
"\xd5\x6c\x03\x77\x0e\xa5\x19\xb0"
"\x47\xde\x52\xe9\x80\x17\x8b\x22"
"\xb9\x2d\xc4\x5b\xf2\x66\xfd\x94"
"\x08\x9f\x36\xcd\x41\xd8\x6f\x06"
"\x7a\x11\xa8\x1c\xb3\x4a\xe1\x55"
"\xec\x83\x1a\x8e\x25\xbc\x30\xc7"
"\x5e\xf5\x69\x00\x97\x0b\xa2\x39"
"\xd0\x44\xdb\x72\x09\x7d\x14\xab"
"\x1f\xb6\x4d\xe4\x58\xef\x86\x1d"
"\x91\x28\xbf\x33\xca\x61\xf8\x6c"
"\x03\x9a\x0e\xa5\x3c\xd3\x47\xde"
"\x75\x0c\x80\x17\xae\x22\xb9\x50"
"\xe7\x5b\xf2\x89\x20\x94\x2b\xc2"
"\x36\xcd\x64\xfb\x6f\x06\x9d\x11"
"\xa8\x3f\xd6\x4a\xe1\x78\x0f\x83"
"\x1a\xb1\x25\xbc\x53\xea\x5e\xf5"
"\x8c\x00\x97\x2e\xc5\x39\xd0\x67"
"\xfe\x72\x09\xa0\x14\xab\x42\xd9"
"\x4d\xe4\x7b\x12\x86\x1d\xb4\x28"
"\xbf\x56\xed\x61\xf8\x8f\x03\x9a"
"\x31\xc8\x3c\xd3\x6a\x01\x75\x0c"
"\xa3\x17\xae\x45\xdc\x50\xe7\x7e"
"\x15\x89\x20\xb7\x2b\xc2\x59\xf0"
"\x64\xfb\x92\x06\x9d\x34\xcb\x3f"
"\xd6\x6d\x04\x78\x0f\xa6\x1a\xb1"
"\x48\xdf\x53\xea\x81\x18\x8c\x23"
"\xba\x2e\xc5\x5c\xf3\x67\xfe\x95"
"\x09\xa0\x37\xce\x42\xd9\x70\x07"
"\x7b\x12\xa9\x1d\xb4\x4b\xe2\x56"
"\xed\x84\x1b\x8f\x26\xbd\x31\xc8"
"\x5f\xf6\x6a\x01\x98\x0c\xa3\x3a"
"\xd1\x45\xdc\x73\x0a\x7e\x15\xac"
"\x20\xb7\x4e\xe5\x59\xf0\x87\x1e"
"\x92\x29\xc0\x34\xcb\x62\xf9\x6d"
"\x04\x9b\x0f\xa6\x3d\xd4\x48\xdf"
"\x76\x0d\x81\x18\xaf\x23\xba\x51"
"\xe8\x5c\xf3\x8a\x21\x95\x2c\xc3"
"\x37\xce\x65\xfc\x70\x07\x9e\x12"
"\xa9\x40\xd7\x4b\xe2\x79\x10\x84"
"\x1b\xb2\x26\xbd\x54\xeb\x5f\xf6"
"\x8d\x01\x98\x2f\xc6\x3a\xd1\x68"
"\xff\x73\x0a\xa1\x15\xac\x43\xda"
"\x4e\xe5\x7c\x13\x87\x1e\xb5\x29"
"\xc0\x57\xee\x62\xf9\x90\x04\x9b"
"\x32\xc9\x3d\xd4\x6b\x02\x76\x0d"
"\xa4\x18\xaf\x46\xdd\x51\xe8\x7f"
"\x16\x8a\x21\xb8\x2c\xc3\x5a\xf1"
"\x65\xfc\x93\x07\x9e\x35\xcc\x40"
"\xd7\x6e\x05\x79\x10\xa7\x1b\xb2"
"\x49\xe0\x54\xeb\x82\x19\x8d\x24"
"\xbb\x2f\xc6\x5d\xf4\x68\xff\x96"
"\x0a\xa1\x38\xcf\x43\xda\x71\x08"
"\x7c\x13\xaa\x1e\xb5\x4c\xe3\x57"
"\xee\x85\x1c\x90\x27\xbe\x32\xc9"
"\x60\xf7\x6b\x02\x99\x0d\xa4\x3b"
"\xd2\x46\xdd\x74\x0b\x7f\x16\xad"
"\x21\xb8\x4f\xe6\x5a\xf1\x88\x1f"
"\x93\x2a\xc1\x35\xcc\x63\xfa\x6e"
"\x05\x9c\x10\xa7\x3e\xd5\x49\xe0"
"\x77\x0e\x82\x19\xb0\x24\xbb\x52"
"\xe9\x5d\xf4\x8b\x22\x96\x2d\xc4"
"\x38\xcf\x66\xfd\x71\x08\x9f\x13"
"\xaa\x41\xd8\x4c\xe3\x7a\x11\x85"
"\x1c\xb3\x27\xbe\x55\xec\x60\xf7"
"\x8e\x02\x99\x30\xc7\x3b\xd2\x69"
"\x00\x74\x0b\xa2\x16\xad\x44\xdb"
"\x4f\xe6\x7d\x14\x88\x1f\xb6\x2a"
"\xc1\x58\xef\x63\xfa\x91\x05\x9c"
"\x33\xca\x3e\xd5\x6c\x03\x77\x0e"
"\xa5\x19\xb0\x47\xde\x52\xe9\x80"
"\x17\x8b\x22\xb9\x2d\xc4\x5b\xf2"
"\x66\xfd\x94\x08\x9f\x36\xcd\x41"
"\xd8\x6f\x06\x7a\x11\xa8\x1c\xb3"
"\x4a\xe1\x55\xec\x83\x1a\x8e\x25"
"\xbc\x30\xc7\x5e\xf5\x69\x00\x97"
"\x0b\xa2\x39\xd0\x44\xdb\x72\x09"
"\x7d\x14\xab\x1f\xb6\x4d\xe4\x58"
"\xef\x86\x1d\x91\x28\xbf\x33\xca"
"\x61\xf8\x6c\x03\x9a\x0e\xa5\x3c"
"\xd3\x47\xde\x75\x0c\x80\x17\xae"
"\x22\xb9\x50\xe7\x5b\xf2\x89\x20"
"\x94\x2b\xc2\x36\xcd\x64\xfb\x6f"
"\x06\x9d\x11\xa8\x3f\xd6\x4a\xe1"
"\x78\x0f\x83\x1a\xb1\x25\xbc\x53"
"\xea\x5e\xf5\x8c\x00\x97\x2e\xc5"
"\x39\xd0\x67\xfe\x72\x09\xa0\x14"
"\xab\x42\xd9\x4d\xe4\x7b\x12\x86"
"\x1d\xb4\x28\xbf\x56\xed\x61\xf8"
"\x8f\x03\x9a\x31\xc8\x3c\xd3\x6a"
"\x01\x75\x0c\xa3\x17\xae\x45\xdc"
"\x50\xe7\x7e\x15\x89\x20\xb7\x2b"
"\xc2\x59\xf0\x64\xfb\x92\x06\x9d"
"\x34\xcb\x3f\xd6\x6d\x04\x78\x0f"
"\xa6\x1a\xb1\x48\xdf\x53\xea\x81"
"\x18\x8c\x23\xba\x2e\xc5\x5c\xf3"
"\x67\xfe\x95\x09\xa0\x37\xce\x42"
"\xd9\x70\x07\x7b\x12\xa9\x1d\xb4"
"\x4b\xe2\x56\xed\x84\x1b\x8f\x26"
"\xbd\x31\xc8\x5f\xf6\x6a\x01\x98",
.psize = 2048,
.digest = "\xfb\x3a\x7a\xda",
}
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
/*
* CRC32C test vectors
*/
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct hash_testvec crc32c_tv_template[] = {
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.psize = 0,
.digest = "\x00\x00\x00\x00",
},
crypto: testmgr - add more unkeyed crc32 and crc32c test vectors crc32c has an unkeyed test vector but crc32 did not. Add the crc32c one (which uses an empty input) to crc32 too, and also add a new one to both that uses a nonempty input. These test vectors verify that crc32 and crc32c implementations use the correct default initial state. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-05-19 23:07:42 -06:00
{
.plaintext = "abcdefg",
.psize = 7,
.digest = "\x41\xf4\x27\xe6",
},
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
{
.key = "\x87\xa9\xcb\xed",
.ksize = 4,
.psize = 0,
.digest = "\x78\x56\x34\x12",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
"\x21\x22\x23\x24\x25\x26\x27\x28",
.psize = 40,
.digest = "\x7f\x15\x2c\x0e",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
.psize = 40,
.digest = "\xf6\xeb\x80\xe9",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78",
.psize = 40,
.digest = "\xed\xbd\x74\xde",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
.psize = 40,
.digest = "\x62\xc8\x79\xd5",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
.psize = 40,
.digest = "\xd0\x9a\x97\xba",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 40,
.digest = "\x13\xd9\x29\x2b",
},
{
.key = "\x80\xea\xd3\xf1",
.ksize = 4,
.plaintext = "\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50",
.psize = 40,
.digest = "\x0c\xb5\xe2\xa2",
},
{
.key = "\xf3\x4a\x1d\x5d",
.ksize = 4,
.plaintext = "\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78",
.psize = 40,
.digest = "\xd1\x7f\xfb\xa6",
},
{
.key = "\x2e\x80\x04\x59",
.ksize = 4,
.plaintext = "\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0",
.psize = 40,
.digest = "\x59\x33\xe6\x7a",
},
{
.key = "\xa6\xcc\x19\x85",
.ksize = 4,
.plaintext = "\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8",
.psize = 40,
.digest = "\xbe\x03\x01\xd2",
},
{
.key = "\x41\xfc\xfe\x2d",
.ksize = 4,
.plaintext = "\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 40,
.digest = "\x75\xd3\xc5\x24",
},
{
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x01\x02\x03\x04\x05\x06\x07\x08"
"\x09\x0a\x0b\x0c\x0d\x0e\x0f\x10"
"\x11\x12\x13\x14\x15\x16\x17\x18"
"\x19\x1a\x1b\x1c\x1d\x1e\x1f\x20"
"\x21\x22\x23\x24\x25\x26\x27\x28"
"\x29\x2a\x2b\x2c\x2d\x2e\x2f\x30"
"\x31\x32\x33\x34\x35\x36\x37\x38"
"\x39\x3a\x3b\x3c\x3d\x3e\x3f\x40"
"\x41\x42\x43\x44\x45\x46\x47\x48"
"\x49\x4a\x4b\x4c\x4d\x4e\x4f\x50"
"\x51\x52\x53\x54\x55\x56\x57\x58"
"\x59\x5a\x5b\x5c\x5d\x5e\x5f\x60"
"\x61\x62\x63\x64\x65\x66\x67\x68"
"\x69\x6a\x6b\x6c\x6d\x6e\x6f\x70"
"\x71\x72\x73\x74\x75\x76\x77\x78"
"\x79\x7a\x7b\x7c\x7d\x7e\x7f\x80"
"\x81\x82\x83\x84\x85\x86\x87\x88"
"\x89\x8a\x8b\x8c\x8d\x8e\x8f\x90"
"\x91\x92\x93\x94\x95\x96\x97\x98"
"\x99\x9a\x9b\x9c\x9d\x9e\x9f\xa0"
"\xa1\xa2\xa3\xa4\xa5\xa6\xa7\xa8"
"\xa9\xaa\xab\xac\xad\xae\xaf\xb0"
"\xb1\xb2\xb3\xb4\xb5\xb6\xb7\xb8"
"\xb9\xba\xbb\xbc\xbd\xbe\xbf\xc0"
"\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8"
"\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0"
"\xd1\xd2\xd3\xd4\xd5\xd6\xd7\xd8"
"\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0"
"\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8"
"\xe9\xea\xeb\xec\xed\xee\xef\xf0",
.psize = 240,
.digest = "\x75\xd3\xc5\x24",
.np = 2,
.tap = { 31, 209 }
crypto: testmgr - add larger crc32c test vector to test FPU path in crc32c_intel Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-11-13 03:02:30 -07:00
}, {
.key = "\xff\xff\xff\xff",
.ksize = 4,
.plaintext = "\x6e\x05\x79\x10\xa7\x1b\xb2\x49"
"\xe0\x54\xeb\x82\x19\x8d\x24\xbb"
"\x2f\xc6\x5d\xf4\x68\xff\x96\x0a"
"\xa1\x38\xcf\x43\xda\x71\x08\x7c"
"\x13\xaa\x1e\xb5\x4c\xe3\x57\xee"
"\x85\x1c\x90\x27\xbe\x32\xc9\x60"
"\xf7\x6b\x02\x99\x0d\xa4\x3b\xd2"
"\x46\xdd\x74\x0b\x7f\x16\xad\x21"
"\xb8\x4f\xe6\x5a\xf1\x88\x1f\x93"
"\x2a\xc1\x35\xcc\x63\xfa\x6e\x05"
"\x9c\x10\xa7\x3e\xd5\x49\xe0\x77"
"\x0e\x82\x19\xb0\x24\xbb\x52\xe9"
"\x5d\xf4\x8b\x22\x96\x2d\xc4\x38"
"\xcf\x66\xfd\x71\x08\x9f\x13\xaa"
"\x41\xd8\x4c\xe3\x7a\x11\x85\x1c"
"\xb3\x27\xbe\x55\xec\x60\xf7\x8e"
"\x02\x99\x30\xc7\x3b\xd2\x69\x00"
"\x74\x0b\xa2\x16\xad\x44\xdb\x4f"
"\xe6\x7d\x14\x88\x1f\xb6\x2a\xc1"
"\x58\xef\x63\xfa\x91\x05\x9c\x33"
"\xca\x3e\xd5\x6c\x03\x77\x0e\xa5"
"\x19\xb0\x47\xde\x52\xe9\x80\x17"
"\x8b\x22\xb9\x2d\xc4\x5b\xf2\x66"
"\xfd\x94\x08\x9f\x36\xcd\x41\xd8"
"\x6f\x06\x7a\x11\xa8\x1c\xb3\x4a"
"\xe1\x55\xec\x83\x1a\x8e\x25\xbc"
"\x30\xc7\x5e\xf5\x69\x00\x97\x0b"
"\xa2\x39\xd0\x44\xdb\x72\x09\x7d"
"\x14\xab\x1f\xb6\x4d\xe4\x58\xef"
"\x86\x1d\x91\x28\xbf\x33\xca\x61"
"\xf8\x6c\x03\x9a\x0e\xa5\x3c\xd3"
"\x47\xde\x75\x0c\x80\x17\xae\x22"
"\xb9\x50\xe7\x5b\xf2\x89\x20\x94"
"\x2b\xc2\x36\xcd\x64\xfb\x6f\x06"
"\x9d\x11\xa8\x3f\xd6\x4a\xe1\x78"
"\x0f\x83\x1a\xb1\x25\xbc\x53\xea"
"\x5e\xf5\x8c\x00\x97\x2e\xc5\x39"
"\xd0\x67\xfe\x72\x09\xa0\x14\xab"
"\x42\xd9\x4d\xe4\x7b\x12\x86\x1d"
"\xb4\x28\xbf\x56\xed\x61\xf8\x8f"
"\x03\x9a\x31\xc8\x3c\xd3\x6a\x01"
"\x75\x0c\xa3\x17\xae\x45\xdc\x50"
"\xe7\x7e\x15\x89\x20\xb7\x2b\xc2"
"\x59\xf0\x64\xfb\x92\x06\x9d\x34"
"\xcb\x3f\xd6\x6d\x04\x78\x0f\xa6"
"\x1a\xb1\x48\xdf\x53\xea\x81\x18"
"\x8c\x23\xba\x2e\xc5\x5c\xf3\x67"
"\xfe\x95\x09\xa0\x37\xce\x42\xd9"
"\x70\x07\x7b\x12\xa9\x1d\xb4\x4b"
"\xe2\x56\xed\x84\x1b\x8f\x26\xbd"
"\x31\xc8\x5f\xf6\x6a\x01\x98\x0c"
"\xa3\x3a\xd1\x45\xdc\x73\x0a\x7e"
"\x15\xac\x20\xb7\x4e\xe5\x59\xf0"
"\x87\x1e\x92\x29\xc0\x34\xcb\x62"
"\xf9\x6d\x04\x9b\x0f\xa6\x3d\xd4"
"\x48\xdf\x76\x0d\x81\x18\xaf\x23"
"\xba\x51\xe8\x5c\xf3\x8a\x21\x95"
"\x2c\xc3\x37\xce\x65\xfc\x70\x07"
"\x9e\x12\xa9\x40\xd7\x4b\xe2\x79"
"\x10\x84\x1b\xb2\x26\xbd\x54\xeb"
"\x5f\xf6\x8d\x01\x98\x2f\xc6\x3a"
"\xd1\x68\xff\x73\x0a\xa1\x15\xac"
"\x43\xda\x4e\xe5\x7c\x13\x87\x1e"
"\xb5\x29\xc0\x57\xee\x62\xf9\x90"
"\x04\x9b\x32\xc9\x3d\xd4\x6b\x02"
"\x76\x0d\xa4\x18\xaf\x46\xdd\x51"
"\xe8\x7f\x16\x8a\x21\xb8\x2c\xc3"
"\x5a\xf1\x65\xfc\x93\x07\x9e\x35"
"\xcc\x40\xd7\x6e\x05\x79\x10\xa7"
"\x1b\xb2\x49\xe0\x54\xeb\x82\x19"
"\x8d\x24\xbb\x2f\xc6\x5d\xf4\x68"
"\xff\x96\x0a\xa1\x38\xcf\x43\xda"
"\x71\x08\x7c\x13\xaa\x1e\xb5\x4c"
"\xe3\x57\xee\x85\x1c\x90\x27\xbe"
"\x32\xc9\x60\xf7\x6b\x02\x99\x0d"
"\xa4\x3b\xd2\x46\xdd\x74\x0b\x7f"
"\x16\xad\x21\xb8\x4f\xe6\x5a\xf1"
"\x88\x1f\x93\x2a\xc1\x35\xcc\x63"
"\xfa\x6e\x05\x9c\x10\xa7\x3e\xd5"
"\x49\xe0\x77\x0e\x82\x19\xb0\x24"
"\xbb\x52\xe9\x5d\xf4\x8b\x22\x96"
"\x2d\xc4\x38\xcf\x66\xfd\x71\x08"
"\x9f\x13\xaa\x41\xd8\x4c\xe3\x7a"
"\x11\x85\x1c\xb3\x27\xbe\x55\xec"
"\x60\xf7\x8e\x02\x99\x30\xc7\x3b"
"\xd2\x69\x00\x74\x0b\xa2\x16\xad"
"\x44\xdb\x4f\xe6\x7d\x14\x88\x1f"
"\xb6\x2a\xc1\x58\xef\x63\xfa\x91"
"\x05\x9c\x33\xca\x3e\xd5\x6c\x03"
"\x77\x0e\xa5\x19\xb0\x47\xde\x52"
"\xe9\x80\x17\x8b\x22\xb9\x2d\xc4"
"\x5b\xf2\x66\xfd\x94\x08\x9f\x36"
"\xcd\x41\xd8\x6f\x06\x7a\x11\xa8"
"\x1c\xb3\x4a\xe1\x55\xec\x83\x1a"
"\x8e\x25\xbc\x30\xc7\x5e\xf5\x69"
"\x00\x97\x0b\xa2\x39\xd0\x44\xdb"
"\x72\x09\x7d\x14\xab\x1f\xb6\x4d"
"\xe4\x58\xef\x86\x1d\x91\x28\xbf"
"\x33\xca\x61\xf8\x6c\x03\x9a\x0e"
"\xa5\x3c\xd3\x47\xde\x75\x0c\x80"
"\x17\xae\x22\xb9\x50\xe7\x5b\xf2"
"\x89\x20\x94\x2b\xc2\x36\xcd\x64"
"\xfb\x6f\x06\x9d\x11\xa8\x3f\xd6"
"\x4a\xe1\x78\x0f\x83\x1a\xb1\x25"
"\xbc\x53\xea\x5e\xf5\x8c\x00\x97"
"\x2e\xc5\x39\xd0\x67\xfe\x72\x09"
"\xa0\x14\xab\x42\xd9\x4d\xe4\x7b"
"\x12\x86\x1d\xb4\x28\xbf\x56\xed"
"\x61\xf8\x8f\x03\x9a\x31\xc8\x3c"
"\xd3\x6a\x01\x75\x0c\xa3\x17\xae"
"\x45\xdc\x50\xe7\x7e\x15\x89\x20"
"\xb7\x2b\xc2\x59\xf0\x64\xfb\x92"
"\x06\x9d\x34\xcb\x3f\xd6\x6d\x04"
"\x78\x0f\xa6\x1a\xb1\x48\xdf\x53"
"\xea\x81\x18\x8c\x23\xba\x2e\xc5"
"\x5c\xf3\x67\xfe\x95\x09\xa0\x37"
"\xce\x42\xd9\x70\x07\x7b\x12\xa9"
"\x1d\xb4\x4b\xe2\x56\xed\x84\x1b"
"\x8f\x26\xbd\x31\xc8\x5f\xf6\x6a"
"\x01\x98\x0c\xa3\x3a\xd1\x45\xdc"
"\x73\x0a\x7e\x15\xac\x20\xb7\x4e"
"\xe5\x59\xf0\x87\x1e\x92\x29\xc0"
"\x34\xcb\x62\xf9\x6d\x04\x9b\x0f"
"\xa6\x3d\xd4\x48\xdf\x76\x0d\x81"
"\x18\xaf\x23\xba\x51\xe8\x5c\xf3"
"\x8a\x21\x95\x2c\xc3\x37\xce\x65"
"\xfc\x70\x07\x9e\x12\xa9\x40\xd7"
"\x4b\xe2\x79\x10\x84\x1b\xb2\x26"
"\xbd\x54\xeb\x5f\xf6\x8d\x01\x98"
"\x2f\xc6\x3a\xd1\x68\xff\x73\x0a"
"\xa1\x15\xac\x43\xda\x4e\xe5\x7c"
"\x13\x87\x1e\xb5\x29\xc0\x57\xee"
"\x62\xf9\x90\x04\x9b\x32\xc9\x3d"
"\xd4\x6b\x02\x76\x0d\xa4\x18\xaf"
"\x46\xdd\x51\xe8\x7f\x16\x8a\x21"
"\xb8\x2c\xc3\x5a\xf1\x65\xfc\x93"
"\x07\x9e\x35\xcc\x40\xd7\x6e\x05"
"\x79\x10\xa7\x1b\xb2\x49\xe0\x54"
"\xeb\x82\x19\x8d\x24\xbb\x2f\xc6"
"\x5d\xf4\x68\xff\x96\x0a\xa1\x38"
"\xcf\x43\xda\x71\x08\x7c\x13\xaa"
"\x1e\xb5\x4c\xe3\x57\xee\x85\x1c"
"\x90\x27\xbe\x32\xc9\x60\xf7\x6b"
"\x02\x99\x0d\xa4\x3b\xd2\x46\xdd"
"\x74\x0b\x7f\x16\xad\x21\xb8\x4f"
"\xe6\x5a\xf1\x88\x1f\x93\x2a\xc1"
"\x35\xcc\x63\xfa\x6e\x05\x9c\x10"
"\xa7\x3e\xd5\x49\xe0\x77\x0e\x82"
"\x19\xb0\x24\xbb\x52\xe9\x5d\xf4"
"\x8b\x22\x96\x2d\xc4\x38\xcf\x66"
"\xfd\x71\x08\x9f\x13\xaa\x41\xd8"
"\x4c\xe3\x7a\x11\x85\x1c\xb3\x27"
"\xbe\x55\xec\x60\xf7\x8e\x02\x99"
"\x30\xc7\x3b\xd2\x69\x00\x74\x0b"
"\xa2\x16\xad\x44\xdb\x4f\xe6\x7d"
"\x14\x88\x1f\xb6\x2a\xc1\x58\xef"
"\x63\xfa\x91\x05\x9c\x33\xca\x3e"
"\xd5\x6c\x03\x77\x0e\xa5\x19\xb0"
"\x47\xde\x52\xe9\x80\x17\x8b\x22"
"\xb9\x2d\xc4\x5b\xf2\x66\xfd\x94"
"\x08\x9f\x36\xcd\x41\xd8\x6f\x06"
"\x7a\x11\xa8\x1c\xb3\x4a\xe1\x55"
"\xec\x83\x1a\x8e\x25\xbc\x30\xc7"
"\x5e\xf5\x69\x00\x97\x0b\xa2\x39"
"\xd0\x44\xdb\x72\x09\x7d\x14\xab"
"\x1f\xb6\x4d\xe4\x58\xef\x86\x1d"
"\x91\x28\xbf\x33\xca\x61\xf8\x6c"
"\x03\x9a\x0e\xa5\x3c\xd3\x47\xde"
"\x75\x0c\x80\x17\xae\x22\xb9\x50"
"\xe7\x5b\xf2\x89\x20\x94\x2b\xc2"
"\x36\xcd\x64\xfb\x6f\x06\x9d\x11"
"\xa8\x3f\xd6\x4a\xe1\x78\x0f\x83"
"\x1a\xb1\x25\xbc\x53\xea\x5e\xf5"
"\x8c\x00\x97\x2e\xc5\x39\xd0\x67"
"\xfe\x72\x09\xa0\x14\xab\x42\xd9"
"\x4d\xe4\x7b\x12\x86\x1d\xb4\x28"
"\xbf\x56\xed\x61\xf8\x8f\x03\x9a"
"\x31\xc8\x3c\xd3\x6a\x01\x75\x0c"
"\xa3\x17\xae\x45\xdc\x50\xe7\x7e"
"\x15\x89\x20\xb7\x2b\xc2\x59\xf0"
"\x64\xfb\x92\x06\x9d\x34\xcb\x3f"
"\xd6\x6d\x04\x78\x0f\xa6\x1a\xb1"
"\x48\xdf\x53\xea\x81\x18\x8c\x23"
"\xba\x2e\xc5\x5c\xf3\x67\xfe\x95"
"\x09\xa0\x37\xce\x42\xd9\x70\x07"
"\x7b\x12\xa9\x1d\xb4\x4b\xe2\x56"
"\xed\x84\x1b\x8f\x26\xbd\x31\xc8"
"\x5f\xf6\x6a\x01\x98\x0c\xa3\x3a"
"\xd1\x45\xdc\x73\x0a\x7e\x15\xac"
"\x20\xb7\x4e\xe5\x59\xf0\x87\x1e"
"\x92\x29\xc0\x34\xcb\x62\xf9\x6d"
"\x04\x9b\x0f\xa6\x3d\xd4\x48\xdf"
"\x76\x0d\x81\x18\xaf\x23\xba\x51"
"\xe8\x5c\xf3\x8a\x21\x95\x2c\xc3"
"\x37\xce\x65\xfc\x70\x07\x9e\x12"
"\xa9\x40\xd7\x4b\xe2\x79\x10\x84"
"\x1b\xb2\x26\xbd\x54\xeb\x5f\xf6"
"\x8d\x01\x98\x2f\xc6\x3a\xd1\x68"
"\xff\x73\x0a\xa1\x15\xac\x43\xda"
"\x4e\xe5\x7c\x13\x87\x1e\xb5\x29"
"\xc0\x57\xee\x62\xf9\x90\x04\x9b"
"\x32\xc9\x3d\xd4\x6b\x02\x76\x0d"
"\xa4\x18\xaf\x46\xdd\x51\xe8\x7f"
"\x16\x8a\x21\xb8\x2c\xc3\x5a\xf1"
"\x65\xfc\x93\x07\x9e\x35\xcc\x40"
"\xd7\x6e\x05\x79\x10\xa7\x1b\xb2"
"\x49\xe0\x54\xeb\x82\x19\x8d\x24"
"\xbb\x2f\xc6\x5d\xf4\x68\xff\x96"
"\x0a\xa1\x38\xcf\x43\xda\x71\x08"
"\x7c\x13\xaa\x1e\xb5\x4c\xe3\x57"
"\xee\x85\x1c\x90\x27\xbe\x32\xc9"
"\x60\xf7\x6b\x02\x99\x0d\xa4\x3b"
"\xd2\x46\xdd\x74\x0b\x7f\x16\xad"
"\x21\xb8\x4f\xe6\x5a\xf1\x88\x1f"
"\x93\x2a\xc1\x35\xcc\x63\xfa\x6e"
"\x05\x9c\x10\xa7\x3e\xd5\x49\xe0"
"\x77\x0e\x82\x19\xb0\x24\xbb\x52"
"\xe9\x5d\xf4\x8b\x22\x96\x2d\xc4"
"\x38\xcf\x66\xfd\x71\x08\x9f\x13"
"\xaa\x41\xd8\x4c\xe3\x7a\x11\x85"
"\x1c\xb3\x27\xbe\x55\xec\x60\xf7"
"\x8e\x02\x99\x30\xc7\x3b\xd2\x69"
"\x00\x74\x0b\xa2\x16\xad\x44\xdb"
"\x4f\xe6\x7d\x14\x88\x1f\xb6\x2a"
"\xc1\x58\xef\x63\xfa\x91\x05\x9c"
"\x33\xca\x3e\xd5\x6c\x03\x77\x0e"
"\xa5\x19\xb0\x47\xde\x52\xe9\x80"
"\x17\x8b\x22\xb9\x2d\xc4\x5b\xf2"
"\x66\xfd\x94\x08\x9f\x36\xcd\x41"
"\xd8\x6f\x06\x7a\x11\xa8\x1c\xb3"
"\x4a\xe1\x55\xec\x83\x1a\x8e\x25"
"\xbc\x30\xc7\x5e\xf5\x69\x00\x97"
"\x0b\xa2\x39\xd0\x44\xdb\x72\x09"
"\x7d\x14\xab\x1f\xb6\x4d\xe4\x58"
"\xef\x86\x1d\x91\x28\xbf\x33\xca"
"\x61\xf8\x6c\x03\x9a\x0e\xa5\x3c"
"\xd3\x47\xde\x75\x0c\x80\x17\xae"
"\x22\xb9\x50\xe7\x5b\xf2\x89\x20"
"\x94\x2b\xc2\x36\xcd\x64\xfb\x6f"
"\x06\x9d\x11\xa8\x3f\xd6\x4a\xe1"
"\x78\x0f\x83\x1a\xb1\x25\xbc\x53"
"\xea\x5e\xf5\x8c\x00\x97\x2e\xc5"
"\x39\xd0\x67\xfe\x72\x09\xa0\x14"
"\xab\x42\xd9\x4d\xe4\x7b\x12\x86"
"\x1d\xb4\x28\xbf\x56\xed\x61\xf8"
"\x8f\x03\x9a\x31\xc8\x3c\xd3\x6a"
"\x01\x75\x0c\xa3\x17\xae\x45\xdc"
"\x50\xe7\x7e\x15\x89\x20\xb7\x2b"
"\xc2\x59\xf0\x64\xfb\x92\x06\x9d"
"\x34\xcb\x3f\xd6\x6d\x04\x78\x0f"
"\xa6\x1a\xb1\x48\xdf\x53\xea\x81"
"\x18\x8c\x23\xba\x2e\xc5\x5c\xf3"
"\x67\xfe\x95\x09\xa0\x37\xce\x42"
"\xd9\x70\x07\x7b\x12\xa9\x1d\xb4"
"\x4b\xe2\x56\xed\x84\x1b\x8f\x26"
"\xbd\x31\xc8\x5f\xf6\x6a\x01\x98",
.psize = 2048,
.digest = "\xec\x26\x4d\x95",
}
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lz4_comp_tv_template[] = {
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
{
crypto: change LZ4 modules to work with new LZ4 module version Update the crypto modules using LZ4 compression as well as the test cases in testmgr.h to work with the new LZ4 module version. Link: http://lkml.kernel.org/r/1486321748-19085-4-git-send-email-4sschmid@informatik.uni-hamburg.de Signed-off-by: Sven Schmidt <4sschmid@informatik.uni-hamburg.de> Cc: Bongkyu Kim <bongkyu.kim@lge.com> Cc: Rui Salvaterra <rsalvaterra@gmail.com> Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: David S. Miller <davem@davemloft.net> Cc: Anton Vorontsov <anton@enomsg.org> Cc: Colin Cross <ccross@android.com> Cc: Kees Cook <keescook@chromium.org> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-24 16:01:19 -07:00
.inlen = 255,
.outlen = 218,
.input = "LZ4 is lossless compression algorithm, providing"
" compression speed at 400 MB/s per core, scalable "
"with multi-cores CPU. It features an extremely fast "
"decoder, with speed in multiple GB/s per core, "
"typically reaching RAM speed limits on multi-core "
"systems.",
.output = "\xf9\x21\x4c\x5a\x34\x20\x69\x73\x20\x6c\x6f\x73\x73"
"\x6c\x65\x73\x73\x20\x63\x6f\x6d\x70\x72\x65\x73\x73"
"\x69\x6f\x6e\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d"
"\x2c\x20\x70\x72\x6f\x76\x69\x64\x69\x6e\x67\x21\x00"
"\xf0\x21\x73\x70\x65\x65\x64\x20\x61\x74\x20\x34\x30"
"\x30\x20\x4d\x42\x2f\x73\x20\x70\x65\x72\x20\x63\x6f"
"\x72\x65\x2c\x20\x73\x63\x61\x6c\x61\x62\x6c\x65\x20"
"\x77\x69\x74\x68\x20\x6d\x75\x6c\x74\x69\x2d\x1a\x00"
"\xf0\x00\x73\x20\x43\x50\x55\x2e\x20\x49\x74\x20\x66"
"\x65\x61\x74\x75\x11\x00\xf2\x0b\x61\x6e\x20\x65\x78"
"\x74\x72\x65\x6d\x65\x6c\x79\x20\x66\x61\x73\x74\x20"
"\x64\x65\x63\x6f\x64\x65\x72\x2c\x3d\x00\x02\x67\x00"
"\x22\x69\x6e\x46\x00\x5a\x70\x6c\x65\x20\x47\x6c\x00"
"\xf0\x00\x74\x79\x70\x69\x63\x61\x6c\x6c\x79\x20\x72"
"\x65\x61\x63\x68\xa7\x00\x33\x52\x41\x4d\x38\x00\x83"
"\x6c\x69\x6d\x69\x74\x73\x20\x6f\x3f\x00\x01\x85\x00"
"\x90\x20\x73\x79\x73\x74\x65\x6d\x73\x2e",
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lz4_decomp_tv_template[] = {
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
{
crypto: change LZ4 modules to work with new LZ4 module version Update the crypto modules using LZ4 compression as well as the test cases in testmgr.h to work with the new LZ4 module version. Link: http://lkml.kernel.org/r/1486321748-19085-4-git-send-email-4sschmid@informatik.uni-hamburg.de Signed-off-by: Sven Schmidt <4sschmid@informatik.uni-hamburg.de> Cc: Bongkyu Kim <bongkyu.kim@lge.com> Cc: Rui Salvaterra <rsalvaterra@gmail.com> Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: David S. Miller <davem@davemloft.net> Cc: Anton Vorontsov <anton@enomsg.org> Cc: Colin Cross <ccross@android.com> Cc: Kees Cook <keescook@chromium.org> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-24 16:01:19 -07:00
.inlen = 218,
.outlen = 255,
.input = "\xf9\x21\x4c\x5a\x34\x20\x69\x73\x20\x6c\x6f\x73\x73"
"\x6c\x65\x73\x73\x20\x63\x6f\x6d\x70\x72\x65\x73\x73"
"\x69\x6f\x6e\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d"
"\x2c\x20\x70\x72\x6f\x76\x69\x64\x69\x6e\x67\x21\x00"
"\xf0\x21\x73\x70\x65\x65\x64\x20\x61\x74\x20\x34\x30"
"\x30\x20\x4d\x42\x2f\x73\x20\x70\x65\x72\x20\x63\x6f"
"\x72\x65\x2c\x20\x73\x63\x61\x6c\x61\x62\x6c\x65\x20"
"\x77\x69\x74\x68\x20\x6d\x75\x6c\x74\x69\x2d\x1a\x00"
"\xf0\x00\x73\x20\x43\x50\x55\x2e\x20\x49\x74\x20\x66"
"\x65\x61\x74\x75\x11\x00\xf2\x0b\x61\x6e\x20\x65\x78"
"\x74\x72\x65\x6d\x65\x6c\x79\x20\x66\x61\x73\x74\x20"
"\x64\x65\x63\x6f\x64\x65\x72\x2c\x3d\x00\x02\x67\x00"
"\x22\x69\x6e\x46\x00\x5a\x70\x6c\x65\x20\x47\x6c\x00"
"\xf0\x00\x74\x79\x70\x69\x63\x61\x6c\x6c\x79\x20\x72"
"\x65\x61\x63\x68\xa7\x00\x33\x52\x41\x4d\x38\x00\x83"
"\x6c\x69\x6d\x69\x74\x73\x20\x6f\x3f\x00\x01\x85\x00"
"\x90\x20\x73\x79\x73\x74\x65\x6d\x73\x2e",
.output = "LZ4 is lossless compression algorithm, providing"
" compression speed at 400 MB/s per core, scalable "
"with multi-cores CPU. It features an extremely fast "
"decoder, with speed in multiple GB/s per core, "
"typically reaching RAM speed limits on multi-core "
"systems.",
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lz4hc_comp_tv_template[] = {
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
{
crypto: change LZ4 modules to work with new LZ4 module version Update the crypto modules using LZ4 compression as well as the test cases in testmgr.h to work with the new LZ4 module version. Link: http://lkml.kernel.org/r/1486321748-19085-4-git-send-email-4sschmid@informatik.uni-hamburg.de Signed-off-by: Sven Schmidt <4sschmid@informatik.uni-hamburg.de> Cc: Bongkyu Kim <bongkyu.kim@lge.com> Cc: Rui Salvaterra <rsalvaterra@gmail.com> Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: David S. Miller <davem@davemloft.net> Cc: Anton Vorontsov <anton@enomsg.org> Cc: Colin Cross <ccross@android.com> Cc: Kees Cook <keescook@chromium.org> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-24 16:01:19 -07:00
.inlen = 255,
.outlen = 216,
.input = "LZ4 is lossless compression algorithm, providing"
" compression speed at 400 MB/s per core, scalable "
"with multi-cores CPU. It features an extremely fast "
"decoder, with speed in multiple GB/s per core, "
"typically reaching RAM speed limits on multi-core "
"systems.",
.output = "\xf9\x21\x4c\x5a\x34\x20\x69\x73\x20\x6c\x6f\x73\x73"
"\x6c\x65\x73\x73\x20\x63\x6f\x6d\x70\x72\x65\x73\x73"
"\x69\x6f\x6e\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d"
"\x2c\x20\x70\x72\x6f\x76\x69\x64\x69\x6e\x67\x21\x00"
"\xf0\x21\x73\x70\x65\x65\x64\x20\x61\x74\x20\x34\x30"
"\x30\x20\x4d\x42\x2f\x73\x20\x70\x65\x72\x20\x63\x6f"
"\x72\x65\x2c\x20\x73\x63\x61\x6c\x61\x62\x6c\x65\x20"
"\x77\x69\x74\x68\x20\x6d\x75\x6c\x74\x69\x2d\x1a\x00"
"\xf0\x00\x73\x20\x43\x50\x55\x2e\x20\x49\x74\x20\x66"
"\x65\x61\x74\x75\x11\x00\xf2\x0b\x61\x6e\x20\x65\x78"
"\x74\x72\x65\x6d\x65\x6c\x79\x20\x66\x61\x73\x74\x20"
"\x64\x65\x63\x6f\x64\x65\x72\x2c\x3d\x00\x02\x67\x00"
"\x22\x69\x6e\x46\x00\x5a\x70\x6c\x65\x20\x47\x6c\x00"
"\xf0\x00\x74\x79\x70\x69\x63\x61\x6c\x6c\x79\x20\x72"
"\x65\x61\x63\x68\xa7\x00\x33\x52\x41\x4d\x38\x00\x97"
"\x6c\x69\x6d\x69\x74\x73\x20\x6f\x6e\x85\x00\x90\x20"
"\x73\x79\x73\x74\x65\x6d\x73\x2e",
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
},
};
crypto: testmgr - constify all test vectors Cryptographic test vectors should never be modified, so constify them to enforce this at both compile-time and run-time. This moves a significant amount of data from .data to .rodata when the crypto tests are enabled. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2017-02-24 16:46:59 -07:00
static const struct comp_testvec lz4hc_decomp_tv_template[] = {
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
{
crypto: change LZ4 modules to work with new LZ4 module version Update the crypto modules using LZ4 compression as well as the test cases in testmgr.h to work with the new LZ4 module version. Link: http://lkml.kernel.org/r/1486321748-19085-4-git-send-email-4sschmid@informatik.uni-hamburg.de Signed-off-by: Sven Schmidt <4sschmid@informatik.uni-hamburg.de> Cc: Bongkyu Kim <bongkyu.kim@lge.com> Cc: Rui Salvaterra <rsalvaterra@gmail.com> Cc: Sergey Senozhatsky <sergey.senozhatsky@gmail.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: David S. Miller <davem@davemloft.net> Cc: Anton Vorontsov <anton@enomsg.org> Cc: Colin Cross <ccross@android.com> Cc: Kees Cook <keescook@chromium.org> Cc: Tony Luck <tony.luck@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-24 16:01:19 -07:00
.inlen = 216,
.outlen = 255,
.input = "\xf9\x21\x4c\x5a\x34\x20\x69\x73\x20\x6c\x6f\x73\x73"
"\x6c\x65\x73\x73\x20\x63\x6f\x6d\x70\x72\x65\x73\x73"
"\x69\x6f\x6e\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d"
"\x2c\x20\x70\x72\x6f\x76\x69\x64\x69\x6e\x67\x21\x00"
"\xf0\x21\x73\x70\x65\x65\x64\x20\x61\x74\x20\x34\x30"
"\x30\x20\x4d\x42\x2f\x73\x20\x70\x65\x72\x20\x63\x6f"
"\x72\x65\x2c\x20\x73\x63\x61\x6c\x61\x62\x6c\x65\x20"
"\x77\x69\x74\x68\x20\x6d\x75\x6c\x74\x69\x2d\x1a\x00"
"\xf0\x00\x73\x20\x43\x50\x55\x2e\x20\x49\x74\x20\x66"
"\x65\x61\x74\x75\x11\x00\xf2\x0b\x61\x6e\x20\x65\x78"
"\x74\x72\x65\x6d\x65\x6c\x79\x20\x66\x61\x73\x74\x20"
"\x64\x65\x63\x6f\x64\x65\x72\x2c\x3d\x00\x02\x67\x00"
"\x22\x69\x6e\x46\x00\x5a\x70\x6c\x65\x20\x47\x6c\x00"
"\xf0\x00\x74\x79\x70\x69\x63\x61\x6c\x6c\x79\x20\x72"
"\x65\x61\x63\x68\xa7\x00\x33\x52\x41\x4d\x38\x00\x97"
"\x6c\x69\x6d\x69\x74\x73\x20\x6f\x6e\x85\x00\x90\x20"
"\x73\x79\x73\x74\x65\x6d\x73\x2e",
.output = "LZ4 is lossless compression algorithm, providing"
" compression speed at 400 MB/s per core, scalable "
"with multi-cores CPU. It features an extremely fast "
"decoder, with speed in multiple GB/s per core, "
"typically reaching RAM speed limits on multi-core "
"systems.",
crypto: testmgr - add test for lz4 and lz4hc This patch adds a simple test vector for the lz4 and lz4hc compression algorithms. Signed-off-by: KOVACS Krisztian <hidden@sch.bme.hu> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2014-08-22 02:44:36 -06:00
},
};
crypto: zstd - Add zstd support Adds zstd support to crypto and scompress. Only supports the default level. Previously we held off on this patch, since there weren't any users. Now zram is ready for zstd support, but depends on CONFIG_CRYPTO_ZSTD, which isn't defined until this patch is in. I also see a patch adding zstd to pstore [0], which depends on crypto zstd. [0] lkml.kernel.org/r/9c9416b2dff19f05fb4c35879aaa83d11ff72c92.1521626182.git.geliangtang@gmail.com Signed-off-by: Nick Terrell <terrelln@fb.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2018-03-30 13:14:53 -06:00
static const struct comp_testvec zstd_comp_tv_template[] = {
{
.inlen = 68,
.outlen = 39,
.input = "The algorithm is zstd. "
"The algorithm is zstd. "
"The algorithm is zstd.",
.output = "\x28\xb5\x2f\xfd\x00\x50\xf5\x00\x00\xb8\x54\x68\x65"
"\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d\x20\x69\x73"
"\x20\x7a\x73\x74\x64\x2e\x20\x01\x00\x55\x73\x36\x01"
,
},
{
.inlen = 244,
.outlen = 151,
.input = "zstd, short for Zstandard, is a fast lossless "
"compression algorithm, targeting real-time "
"compression scenarios at zlib-level and better "
"compression ratios. The zstd compression library "
"provides in-memory compression and decompression "
"functions.",
.output = "\x28\xb5\x2f\xfd\x00\x50\x75\x04\x00\x42\x4b\x1e\x17"
"\x90\x81\x31\x00\xf2\x2f\xe4\x36\xc9\xef\x92\x88\x32"
"\xc9\xf2\x24\x94\xd8\x68\x9a\x0f\x00\x0c\xc4\x31\x6f"
"\x0d\x0c\x38\xac\x5c\x48\x03\xcd\x63\x67\xc0\xf3\xad"
"\x4e\x90\xaa\x78\xa0\xa4\xc5\x99\xda\x2f\xb6\x24\x60"
"\xe2\x79\x4b\xaa\xb6\x6b\x85\x0b\xc9\xc6\x04\x66\x86"
"\xe2\xcc\xe2\x25\x3f\x4f\x09\xcd\xb8\x9d\xdb\xc1\x90"
"\xa9\x11\xbc\x35\x44\x69\x2d\x9c\x64\x4f\x13\x31\x64"
"\xcc\xfb\x4d\x95\x93\x86\x7f\x33\x7f\x1a\xef\xe9\x30"
"\xf9\x67\xa1\x94\x0a\x69\x0f\x60\xcd\xc3\xab\x99\xdc"
"\x42\xed\x97\x05\x00\x33\xc3\x15\x95\x3a\x06\xa0\x0e"
"\x20\xa9\x0e\x82\xb9\x43\x45\x01",
},
};
static const struct comp_testvec zstd_decomp_tv_template[] = {
{
.inlen = 43,
.outlen = 68,
.input = "\x28\xb5\x2f\xfd\x04\x50\xf5\x00\x00\xb8\x54\x68\x65"
"\x20\x61\x6c\x67\x6f\x72\x69\x74\x68\x6d\x20\x69\x73"
"\x20\x7a\x73\x74\x64\x2e\x20\x01\x00\x55\x73\x36\x01"
"\x6b\xf4\x13\x35",
.output = "The algorithm is zstd. "
"The algorithm is zstd. "
"The algorithm is zstd.",
},
{
.inlen = 155,
.outlen = 244,
.input = "\x28\xb5\x2f\xfd\x04\x50\x75\x04\x00\x42\x4b\x1e\x17"
"\x90\x81\x31\x00\xf2\x2f\xe4\x36\xc9\xef\x92\x88\x32"
"\xc9\xf2\x24\x94\xd8\x68\x9a\x0f\x00\x0c\xc4\x31\x6f"
"\x0d\x0c\x38\xac\x5c\x48\x03\xcd\x63\x67\xc0\xf3\xad"
"\x4e\x90\xaa\x78\xa0\xa4\xc5\x99\xda\x2f\xb6\x24\x60"
"\xe2\x79\x4b\xaa\xb6\x6b\x85\x0b\xc9\xc6\x04\x66\x86"
"\xe2\xcc\xe2\x25\x3f\x4f\x09\xcd\xb8\x9d\xdb\xc1\x90"
"\xa9\x11\xbc\x35\x44\x69\x2d\x9c\x64\x4f\x13\x31\x64"
"\xcc\xfb\x4d\x95\x93\x86\x7f\x33\x7f\x1a\xef\xe9\x30"
"\xf9\x67\xa1\x94\x0a\x69\x0f\x60\xcd\xc3\xab\x99\xdc"
"\x42\xed\x97\x05\x00\x33\xc3\x15\x95\x3a\x06\xa0\x0e"
"\x20\xa9\x0e\x82\xb9\x43\x45\x01\xaa\x6d\xda\x0d",
.output = "zstd, short for Zstandard, is a fast lossless "
"compression algorithm, targeting real-time "
"compression scenarios at zlib-level and better "
"compression ratios. The zstd compression library "
"provides in-memory compression and decompression "
"functions.",
},
};
crypto: cryptomgr - Add test infrastructure This patch moves the newly created alg_test infrastructure into cryptomgr. This shall allow us to use it for testing at algorithm registrations. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2008-07-31 03:08:25 -06:00
#endif /* _CRYPTO_TESTMGR_H */
Reference in a new issue Copy permalink