commit 4a8108b70508df0b6c4ffa4a3974dab93dcbe851 upstream.
If the user-provided IV needs to be aligned to the algorithm's
alignmask, then skcipher_walk_virt() copies the IV into a new aligned
buffer walk.iv. But skcipher_walk_virt() can fail afterwards, and then
if the caller unconditionally accesses walk.iv, it's a use-after-free.
xts-aes-neonbs doesn't set an alignmask, so currently it isn't affected
by this despite unconditionally accessing walk.iv. However this is more
subtle than desired, and unconditionally accessing walk.iv has caused a
real problem in other algorithms. Thus, update xts-aes-neonbs to start
checking the return value of skcipher_walk_virt().
Fixes: 1abee99eaf ("crypto: arm64/aes - reimplement bit-sliced ARM/NEON implementation for arm64")
Cc: <stable@vger.kernel.org> # v4.11+
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When kernel mode NEON was first introduced on arm64, the preserve and
restore of the userland NEON state was completely unoptimized, and
involved saving all registers on each call to kernel_neon_begin(),
and restoring them on each call to kernel_neon_end(). For this reason,
the NEON crypto code that was introduced at the time keeps the NEON
enabled throughout the execution of the crypto API methods, which may
include calls back into the crypto API that could result in memory
allocation or other actions that we should avoid when running with
preemption disabled.
Since then, we have optimized the kernel mode NEON handling, which now
restores lazily (upon return to userland), and so the preserve action
is only costly the first time it is called after entering the kernel.
So let's put the kernel_neon_begin() and kernel_neon_end() calls around
the actual invocations of the NEON crypto code, and run the remainder of
the code with kernel mode NEON disabled (and preemption enabled)
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When kernel mode NEON was first introduced on arm64, the preserve and
restore of the userland NEON state was completely unoptimized, and
involved saving all registers on each call to kernel_neon_begin(),
and restoring them on each call to kernel_neon_end(). For this reason,
the NEON crypto code that was introduced at the time keeps the NEON
enabled throughout the execution of the crypto API methods, which may
include calls back into the crypto API that could result in memory
allocation or other actions that we should avoid when running with
preemption disabled.
Since then, we have optimized the kernel mode NEON handling, which now
restores lazily (upon return to userland), and so the preserve action
is only costly the first time it is called after entering the kernel.
So let's put the kernel_neon_begin() and kernel_neon_end() calls around
the actual invocations of the NEON crypto code, and run the remainder of
the code with kernel mode NEON disabled (and preemption enabled)
Note that this requires some reshuffling of the registers in the asm
code, because the XTS routines can no longer rely on the registers to
retain their contents between invocations.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Of the various chaining modes implemented by the bit sliced AES driver,
only CTR is exposed as a synchronous cipher, and requires a fallback in
order to remain usable once we update the kernel mode NEON handling logic
to disallow nested use. So wire up the existing CTR fallback C code.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
There are quite a number of occurrences in the kernel of the pattern
if (dst != src)
memcpy(dst, src, walk.total % AES_BLOCK_SIZE);
crypto_xor(dst, final, walk.total % AES_BLOCK_SIZE);
or
crypto_xor(keystream, src, nbytes);
memcpy(dst, keystream, nbytes);
where crypto_xor() is preceded or followed by a memcpy() invocation
that is only there because crypto_xor() uses its output parameter as
one of the inputs. To avoid having to add new instances of this pattern
in the arm64 code, which will be refactored to implement non-SIMD
fallbacks, add an alternative implementation called crypto_xor_cpy(),
taking separate input and output arguments. This removes the need for
the separate memcpy().
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The arm64 bit sliced AES core code uses the IV buffer to pass the final
keystream block back to the glue code if the input is not a multiple of
the block size, so that the asm code does not have to deal with anything
except 16 byte blocks. This is done under the assumption that the outgoing
IV is meaningless anyway in this case, given that chaining is no longer
possible under these circumstances.
However, as it turns out, the CCM driver does expect the IV to retain
a value that is equal to the original IV except for the counter value,
and even interprets byte zero as a length indicator, which may result
in memory corruption if the IV is overwritten with something else.
So use a separate buffer to return the final keystream block.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The new bitsliced NEON implementation of AES uses a fallback in two
places: CBC encryption (which is strictly sequential, whereas this
driver can only operate efficiently on 8 blocks at a time), and the
XTS tweak generation, which involves encrypting a single AES block
with a different key schedule.
The plain (i.e., non-bitsliced) NEON code is more suitable as a fallback,
given that it is faster than scalar on low end cores (which is what
the NEON implementations target, since high end cores have dedicated
instructions for AES), and shows similar behavior in terms of D-cache
footprint and sensitivity to cache timing attacks. So switch the fallback
handling to the plain NEON driver.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This is a reimplementation of the NEON version of the bit-sliced AES
algorithm. This code is heavily based on Andy Polyakov's OpenSSL version
for ARM, which is also available in the kernel. This is an alternative for
the existing NEON implementation for arm64 authored by me, which suffers
from poor performance due to its reliance on the pathologically slow four
register variant of the tbl/tbx NEON instruction.
This version is about ~30% (*) faster than the generic C code, but only in
cases where the input can be 8x interleaved (this is a fundamental property
of bit slicing). For this reason, only the chaining modes ECB, XTS and CTR
are implemented. (The significance of ECB is that it could potentially be
used by other chaining modes)
* Measured on Cortex-A57. Note that this is still an order of magnitude
slower than the implementations that use the dedicated AES instructions
introduced in ARMv8, but those are part of an optional extension, and so
it is good to have a fallback.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
