2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* INET An implementation of the TCP/IP protocol suite for the LINUX
|
|
|
|
* operating system. INET is implemented using the BSD Socket
|
|
|
|
* interface as the means of communication with the user level.
|
|
|
|
*
|
|
|
|
* PACKET - implements raw packet sockets.
|
|
|
|
*
|
2005-05-05 17:16:16 -06:00
|
|
|
* Authors: Ross Biro
|
2005-04-16 16:20:36 -06:00
|
|
|
* Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
|
|
|
|
* Alan Cox, <gw4pts@gw4pts.ampr.org>
|
|
|
|
*
|
2007-02-09 07:25:10 -07:00
|
|
|
* Fixes:
|
2005-04-16 16:20:36 -06:00
|
|
|
* Alan Cox : verify_area() now used correctly
|
|
|
|
* Alan Cox : new skbuff lists, look ma no backlogs!
|
|
|
|
* Alan Cox : tidied skbuff lists.
|
|
|
|
* Alan Cox : Now uses generic datagram routines I
|
|
|
|
* added. Also fixed the peek/read crash
|
|
|
|
* from all old Linux datagram code.
|
|
|
|
* Alan Cox : Uses the improved datagram code.
|
|
|
|
* Alan Cox : Added NULL's for socket options.
|
|
|
|
* Alan Cox : Re-commented the code.
|
|
|
|
* Alan Cox : Use new kernel side addressing
|
|
|
|
* Rob Janssen : Correct MTU usage.
|
|
|
|
* Dave Platt : Counter leaks caused by incorrect
|
|
|
|
* interrupt locking and some slightly
|
|
|
|
* dubious gcc output. Can you read
|
|
|
|
* compiler: it said _VOLATILE_
|
|
|
|
* Richard Kooijman : Timestamp fixes.
|
|
|
|
* Alan Cox : New buffers. Use sk->mac.raw.
|
|
|
|
* Alan Cox : sendmsg/recvmsg support.
|
|
|
|
* Alan Cox : Protocol setting support
|
|
|
|
* Alexey Kuznetsov : Untied from IPv4 stack.
|
|
|
|
* Cyrus Durgin : Fixed kerneld for kmod.
|
|
|
|
* Michal Ostrowski : Module initialization cleanup.
|
2007-02-09 07:25:10 -07:00
|
|
|
* Ulises Alonso : Frame number limit removal and
|
2005-04-16 16:20:36 -06:00
|
|
|
* packet_set_ring memory leak.
|
2005-09-21 01:11:37 -06:00
|
|
|
* Eric Biederman : Allow for > 8 byte hardware addresses.
|
|
|
|
* The convention is that longer addresses
|
|
|
|
* will simply extend the hardware address
|
2007-02-09 07:25:10 -07:00
|
|
|
* byte arrays at the end of sockaddr_ll
|
2005-09-21 01:11:37 -06:00
|
|
|
* and packet_mreq.
|
2009-05-18 23:11:22 -06:00
|
|
|
* Johann Baudy : Added TX RING.
|
2011-08-19 04:18:16 -06:00
|
|
|
* Chetan Loke : Implemented TPACKET_V3 block abstraction
|
|
|
|
* layer.
|
|
|
|
* Copyright (C) 2011, <lokec@ccs.neu.edu>
|
|
|
|
*
|
2005-04-16 16:20:36 -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.
|
|
|
|
*
|
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/mm.h>
|
2006-01-11 13:17:47 -07:00
|
|
|
#include <linux/capability.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/fcntl.h>
|
|
|
|
#include <linux/socket.h>
|
|
|
|
#include <linux/in.h>
|
|
|
|
#include <linux/inet.h>
|
|
|
|
#include <linux/netdevice.h>
|
|
|
|
#include <linux/if_packet.h>
|
|
|
|
#include <linux/wireless.h>
|
2007-02-05 00:33:10 -07:00
|
|
|
#include <linux/kernel.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/kmod.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 02:04:11 -06:00
|
|
|
#include <linux/slab.h>
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
#include <linux/vmalloc.h>
|
2007-09-12 04:01:34 -06:00
|
|
|
#include <net/net_namespace.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <net/ip.h>
|
|
|
|
#include <net/protocol.h>
|
|
|
|
#include <linux/skbuff.h>
|
|
|
|
#include <net/sock.h>
|
|
|
|
#include <linux/errno.h>
|
|
|
|
#include <linux/timer.h>
|
|
|
|
#include <asm/system.h>
|
|
|
|
#include <asm/uaccess.h>
|
|
|
|
#include <asm/ioctls.h>
|
|
|
|
#include <asm/page.h>
|
2006-10-19 14:08:53 -06:00
|
|
|
#include <asm/cacheflush.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <asm/io.h>
|
|
|
|
#include <linux/proc_fs.h>
|
|
|
|
#include <linux/seq_file.h>
|
|
|
|
#include <linux/poll.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
2009-01-30 15:12:06 -07:00
|
|
|
#include <linux/mutex.h>
|
2009-10-26 19:40:35 -06:00
|
|
|
#include <linux/if_vlan.h>
|
2010-02-04 21:24:10 -07:00
|
|
|
#include <linux/virtio_net.h>
|
2010-04-07 16:41:28 -06:00
|
|
|
#include <linux/errqueue.h>
|
2010-06-02 06:53:56 -06:00
|
|
|
#include <linux/net_tstamp.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
#ifdef CONFIG_INET
|
|
|
|
#include <net/inet_common.h>
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/*
|
|
|
|
Assumptions:
|
|
|
|
- if device has no dev->hard_header routine, it adds and removes ll header
|
|
|
|
inside itself. In this case ll header is invisible outside of device,
|
|
|
|
but higher levels still should reserve dev->hard_header_len.
|
|
|
|
Some devices are enough clever to reallocate skb, when header
|
|
|
|
will not fit to reserved space (tunnel), another ones are silly
|
|
|
|
(PPP).
|
|
|
|
- packet socket receives packets with pulled ll header,
|
|
|
|
so that SOCK_RAW should push it back.
|
|
|
|
|
|
|
|
On receive:
|
|
|
|
-----------
|
|
|
|
|
|
|
|
Incoming, dev->hard_header!=NULL
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> ll header
|
|
|
|
data -> data
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
Outgoing, dev->hard_header!=NULL
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> ll header
|
|
|
|
data -> ll header
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
Incoming, dev->hard_header==NULL
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> UNKNOWN position. It is very likely, that it points to ll
|
|
|
|
header. PPP makes it, that is wrong, because introduce
|
2007-07-18 19:44:35 -06:00
|
|
|
assymetry between rx and tx paths.
|
2007-04-10 22:21:55 -06:00
|
|
|
data -> data
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
Outgoing, dev->hard_header==NULL
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> data. ll header is still not built!
|
|
|
|
data -> data
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
Resume
|
|
|
|
If dev->hard_header==NULL we are unlikely to restore sensible ll header.
|
|
|
|
|
|
|
|
|
|
|
|
On transmit:
|
|
|
|
------------
|
|
|
|
|
|
|
|
dev->hard_header != NULL
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> ll header
|
|
|
|
data -> ll header
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
dev->hard_header == NULL (ll header is added by device, we cannot control it)
|
2007-04-10 22:21:55 -06:00
|
|
|
mac_header -> data
|
|
|
|
data -> data
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
We should set nh.raw on output to correct posistion,
|
|
|
|
packet classifier depends on it.
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* Private packet socket structures. */
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
struct packet_mclist {
|
2005-04-16 16:20:36 -06:00
|
|
|
struct packet_mclist *next;
|
|
|
|
int ifindex;
|
|
|
|
int count;
|
|
|
|
unsigned short type;
|
|
|
|
unsigned short alen;
|
2005-09-21 01:11:37 -06:00
|
|
|
unsigned char addr[MAX_ADDR_LEN];
|
|
|
|
};
|
|
|
|
/* identical to struct packet_mreq except it has
|
|
|
|
* a longer address field.
|
|
|
|
*/
|
2009-07-21 15:57:59 -06:00
|
|
|
struct packet_mreq_max {
|
2005-09-21 01:11:37 -06:00
|
|
|
int mr_ifindex;
|
|
|
|
unsigned short mr_type;
|
|
|
|
unsigned short mr_alen;
|
|
|
|
unsigned char mr_address[MAX_ADDR_LEN];
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
2007-05-29 14:12:50 -06:00
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
|
2009-05-18 23:11:22 -06:00
|
|
|
int closing, int tx_ring);
|
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
#define V3_ALIGNMENT (8)
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
#define BLK_HDR_LEN (ALIGN(sizeof(struct tpacket_block_desc), V3_ALIGNMENT))
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
#define BLK_PLUS_PRIV(sz_of_priv) \
|
|
|
|
(BLK_HDR_LEN + ALIGN((sz_of_priv), V3_ALIGNMENT))
|
|
|
|
|
|
|
|
/* kbdq - kernel block descriptor queue */
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core {
|
2011-08-19 04:18:16 -06:00
|
|
|
struct pgv *pkbdq;
|
|
|
|
unsigned int feature_req_word;
|
|
|
|
unsigned int hdrlen;
|
|
|
|
unsigned char reset_pending_on_curr_blk;
|
|
|
|
unsigned char delete_blk_timer;
|
|
|
|
unsigned short kactive_blk_num;
|
|
|
|
unsigned short blk_sizeof_priv;
|
|
|
|
|
|
|
|
/* last_kactive_blk_num:
|
|
|
|
* trick to see if user-space has caught up
|
|
|
|
* in order to avoid refreshing timer when every single pkt arrives.
|
|
|
|
*/
|
|
|
|
unsigned short last_kactive_blk_num;
|
|
|
|
|
|
|
|
char *pkblk_start;
|
|
|
|
char *pkblk_end;
|
|
|
|
int kblk_size;
|
|
|
|
unsigned int knum_blocks;
|
|
|
|
uint64_t knxt_seq_num;
|
|
|
|
char *prev;
|
|
|
|
char *nxt_offset;
|
|
|
|
struct sk_buff *skb;
|
|
|
|
|
|
|
|
atomic_t blk_fill_in_prog;
|
|
|
|
|
|
|
|
/* Default is set to 8ms */
|
|
|
|
#define DEFAULT_PRB_RETIRE_TOV (8)
|
|
|
|
|
|
|
|
unsigned short retire_blk_tov;
|
|
|
|
unsigned short version;
|
|
|
|
unsigned long tov_in_jiffies;
|
|
|
|
|
|
|
|
/* timer to retire an outstanding block */
|
|
|
|
struct timer_list retire_blk_timer;
|
|
|
|
};
|
|
|
|
|
|
|
|
#define PGV_FROM_VMALLOC 1
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
struct pgv {
|
|
|
|
char *buffer;
|
|
|
|
};
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_ring_buffer {
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
struct pgv *pg_vec;
|
2009-05-18 23:11:22 -06:00
|
|
|
unsigned int head;
|
|
|
|
unsigned int frames_per_block;
|
|
|
|
unsigned int frame_size;
|
|
|
|
unsigned int frame_max;
|
|
|
|
|
|
|
|
unsigned int pg_vec_order;
|
|
|
|
unsigned int pg_vec_pages;
|
|
|
|
unsigned int pg_vec_len;
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core prb_bdqc;
|
2009-05-18 23:11:22 -06:00
|
|
|
atomic_t pending;
|
|
|
|
};
|
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
#define BLOCK_STATUS(x) ((x)->hdr.bh1.block_status)
|
|
|
|
#define BLOCK_NUM_PKTS(x) ((x)->hdr.bh1.num_pkts)
|
|
|
|
#define BLOCK_O2FP(x) ((x)->hdr.bh1.offset_to_first_pkt)
|
|
|
|
#define BLOCK_LEN(x) ((x)->hdr.bh1.blk_len)
|
|
|
|
#define BLOCK_SNUM(x) ((x)->hdr.bh1.seq_num)
|
|
|
|
#define BLOCK_O2PRIV(x) ((x)->offset_to_priv)
|
|
|
|
#define BLOCK_PRIV(x) ((void *)((char *)(x) + BLOCK_O2PRIV(x)))
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_sock;
|
|
|
|
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
static void *packet_previous_frame(struct packet_sock *po,
|
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
int status);
|
|
|
|
static void packet_increment_head(struct packet_ring_buffer *buff);
|
2011-08-25 04:43:30 -06:00
|
|
|
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *,
|
|
|
|
struct tpacket_block_desc *);
|
|
|
|
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *);
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_retire_current_block(struct tpacket_kbdq_core *,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *, unsigned int status);
|
2011-08-25 04:43:30 -06:00
|
|
|
static int prb_queue_frozen(struct tpacket_kbdq_core *);
|
|
|
|
static void prb_open_block(struct tpacket_kbdq_core *,
|
|
|
|
struct tpacket_block_desc *);
|
2011-08-19 04:18:16 -06:00
|
|
|
static void prb_retire_rx_blk_timer_expired(unsigned long);
|
2011-08-25 04:43:30 -06:00
|
|
|
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *);
|
|
|
|
static void prb_init_blk_timer(struct packet_sock *,
|
|
|
|
struct tpacket_kbdq_core *,
|
|
|
|
void (*func) (unsigned long));
|
|
|
|
static void prb_fill_rxhash(struct tpacket_kbdq_core *, struct tpacket3_hdr *);
|
|
|
|
static void prb_clear_rxhash(struct tpacket_kbdq_core *,
|
|
|
|
struct tpacket3_hdr *);
|
|
|
|
static void prb_fill_vlan_info(struct tpacket_kbdq_core *,
|
|
|
|
struct tpacket3_hdr *);
|
2005-04-16 16:20:36 -06:00
|
|
|
static void packet_flush_mclist(struct sock *sk);
|
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
struct packet_fanout;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct packet_sock {
|
|
|
|
/* struct sock has to be the first member of packet_sock */
|
|
|
|
struct sock sk;
|
2011-07-05 02:45:05 -06:00
|
|
|
struct packet_fanout *fanout;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct tpacket_stats stats;
|
2011-08-19 04:18:16 -06:00
|
|
|
union tpacket_stats_u stats_u;
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_ring_buffer rx_ring;
|
|
|
|
struct packet_ring_buffer tx_ring;
|
2005-04-16 16:20:36 -06:00
|
|
|
int copy_thresh;
|
|
|
|
spinlock_t bind_lock;
|
2009-01-30 15:12:06 -07:00
|
|
|
struct mutex pg_vec_lock;
|
2007-02-05 00:31:32 -07:00
|
|
|
unsigned int running:1, /* prot_hook is attached*/
|
2007-04-20 17:05:39 -06:00
|
|
|
auxdata:1,
|
2010-02-04 21:24:10 -07:00
|
|
|
origdev:1,
|
|
|
|
has_vnet_hdr:1;
|
2005-04-16 16:20:36 -06:00
|
|
|
int ifindex; /* bound device */
|
2006-11-08 01:26:29 -07:00
|
|
|
__be16 num;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct packet_mclist *mclist;
|
|
|
|
atomic_t mapped;
|
2008-07-14 23:50:15 -06:00
|
|
|
enum tpacket_versions tp_version;
|
|
|
|
unsigned int tp_hdrlen;
|
2008-07-18 19:05:19 -06:00
|
|
|
unsigned int tp_reserve;
|
2009-05-18 23:11:22 -06:00
|
|
|
unsigned int tp_loss:1;
|
2010-06-02 06:53:56 -06:00
|
|
|
unsigned int tp_tstamp;
|
2009-10-15 22:02:20 -06:00
|
|
|
struct packet_type prot_hook ____cacheline_aligned_in_smp;
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
#define PACKET_FANOUT_MAX 256
|
|
|
|
|
|
|
|
struct packet_fanout {
|
|
|
|
#ifdef CONFIG_NET_NS
|
|
|
|
struct net *net;
|
|
|
|
#endif
|
|
|
|
unsigned int num_members;
|
|
|
|
u16 id;
|
|
|
|
u8 type;
|
2011-07-05 02:43:20 -06:00
|
|
|
u8 defrag;
|
2011-07-05 02:45:05 -06:00
|
|
|
atomic_t rr_cur;
|
|
|
|
struct list_head list;
|
|
|
|
struct sock *arr[PACKET_FANOUT_MAX];
|
|
|
|
spinlock_t lock;
|
|
|
|
atomic_t sk_ref;
|
|
|
|
struct packet_type prot_hook ____cacheline_aligned_in_smp;
|
|
|
|
};
|
|
|
|
|
2007-02-05 00:33:10 -07:00
|
|
|
struct packet_skb_cb {
|
|
|
|
unsigned int origlen;
|
|
|
|
union {
|
|
|
|
struct sockaddr_pkt pkt;
|
|
|
|
struct sockaddr_ll ll;
|
|
|
|
} sa;
|
|
|
|
};
|
|
|
|
|
|
|
|
#define PACKET_SKB_CB(__skb) ((struct packet_skb_cb *)((__skb)->cb))
|
2007-02-05 00:31:32 -07:00
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
#define GET_PBDQC_FROM_RB(x) ((struct tpacket_kbdq_core *)(&(x)->prb_bdqc))
|
2011-08-19 04:18:16 -06:00
|
|
|
#define GET_PBLOCK_DESC(x, bid) \
|
2011-08-25 04:43:30 -06:00
|
|
|
((struct tpacket_block_desc *)((x)->pkbdq[(bid)].buffer))
|
2011-08-19 04:18:16 -06:00
|
|
|
#define GET_CURR_PBLOCK_DESC_FROM_CORE(x) \
|
2011-08-25 04:43:30 -06:00
|
|
|
((struct tpacket_block_desc *)((x)->pkbdq[(x)->kactive_blk_num].buffer))
|
2011-08-19 04:18:16 -06:00
|
|
|
#define GET_NEXT_PRB_BLK_NUM(x) \
|
|
|
|
(((x)->kactive_blk_num < ((x)->knum_blocks-1)) ? \
|
|
|
|
((x)->kactive_blk_num+1) : 0)
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static struct packet_sock *pkt_sk(struct sock *sk)
|
2011-07-04 02:44:29 -06:00
|
|
|
{
|
|
|
|
return (struct packet_sock *)sk;
|
|
|
|
}
|
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
static void __fanout_unlink(struct sock *sk, struct packet_sock *po);
|
|
|
|
static void __fanout_link(struct sock *sk, struct packet_sock *po);
|
|
|
|
|
2011-07-04 02:44:29 -06:00
|
|
|
/* register_prot_hook must be invoked with the po->bind_lock held,
|
|
|
|
* or from a context in which asynchronous accesses to the packet
|
|
|
|
* socket is not possible (packet_create()).
|
|
|
|
*/
|
|
|
|
static void register_prot_hook(struct sock *sk)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
if (!po->running) {
|
2011-07-05 02:45:05 -06:00
|
|
|
if (po->fanout)
|
|
|
|
__fanout_link(sk, po);
|
|
|
|
else
|
|
|
|
dev_add_pack(&po->prot_hook);
|
2011-07-04 02:44:29 -06:00
|
|
|
sock_hold(sk);
|
|
|
|
po->running = 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* {,__}unregister_prot_hook() must be invoked with the po->bind_lock
|
|
|
|
* held. If the sync parameter is true, we will temporarily drop
|
|
|
|
* the po->bind_lock and do a synchronize_net to make sure no
|
|
|
|
* asynchronous packet processing paths still refer to the elements
|
|
|
|
* of po->prot_hook. If the sync parameter is false, it is the
|
|
|
|
* callers responsibility to take care of this.
|
|
|
|
*/
|
|
|
|
static void __unregister_prot_hook(struct sock *sk, bool sync)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
|
|
|
|
po->running = 0;
|
2011-07-05 02:45:05 -06:00
|
|
|
if (po->fanout)
|
|
|
|
__fanout_unlink(sk, po);
|
|
|
|
else
|
|
|
|
__dev_remove_pack(&po->prot_hook);
|
2011-07-04 02:44:29 -06:00
|
|
|
__sock_put(sk);
|
|
|
|
|
|
|
|
if (sync) {
|
|
|
|
spin_unlock(&po->bind_lock);
|
|
|
|
synchronize_net();
|
|
|
|
spin_lock(&po->bind_lock);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void unregister_prot_hook(struct sock *sk, bool sync)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
|
|
|
|
if (po->running)
|
|
|
|
__unregister_prot_hook(sk, sync);
|
|
|
|
}
|
|
|
|
|
2010-12-06 21:26:16 -07:00
|
|
|
static inline __pure struct page *pgv_to_page(void *addr)
|
2010-11-30 19:52:20 -07:00
|
|
|
{
|
|
|
|
if (is_vmalloc_addr(addr))
|
|
|
|
return vmalloc_to_page(addr);
|
|
|
|
return virt_to_page(addr);
|
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static void __packet_set_status(struct packet_sock *po, void *frame, int status)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2008-07-14 23:50:15 -06:00
|
|
|
union {
|
|
|
|
struct tpacket_hdr *h1;
|
|
|
|
struct tpacket2_hdr *h2;
|
|
|
|
void *raw;
|
|
|
|
} h;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
h.raw = frame;
|
2008-07-14 23:50:15 -06:00
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
2009-05-18 23:11:22 -06:00
|
|
|
h.h1->tp_status = status;
|
2010-11-30 19:52:20 -07:00
|
|
|
flush_dcache_page(pgv_to_page(&h.h1->tp_status));
|
2008-07-14 23:50:15 -06:00
|
|
|
break;
|
|
|
|
case TPACKET_V2:
|
2009-05-18 23:11:22 -06:00
|
|
|
h.h2->tp_status = status;
|
2010-11-30 19:52:20 -07:00
|
|
|
flush_dcache_page(pgv_to_page(&h.h2->tp_status));
|
2008-07-14 23:50:15 -06:00
|
|
|
break;
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
2009-05-18 23:11:22 -06:00
|
|
|
default:
|
2011-08-19 04:18:16 -06:00
|
|
|
WARN(1, "TPACKET version not supported.\n");
|
2009-05-18 23:11:22 -06:00
|
|
|
BUG();
|
2008-07-14 23:50:15 -06:00
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
smp_wmb();
|
2008-07-14 23:50:15 -06:00
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static int __packet_get_status(struct packet_sock *po, void *frame)
|
2008-07-14 23:50:15 -06:00
|
|
|
{
|
|
|
|
union {
|
|
|
|
struct tpacket_hdr *h1;
|
|
|
|
struct tpacket2_hdr *h2;
|
|
|
|
void *raw;
|
|
|
|
} h;
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
smp_rmb();
|
|
|
|
|
2008-07-14 23:50:15 -06:00
|
|
|
h.raw = frame;
|
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
2010-11-30 19:52:20 -07:00
|
|
|
flush_dcache_page(pgv_to_page(&h.h1->tp_status));
|
2009-05-18 23:11:22 -06:00
|
|
|
return h.h1->tp_status;
|
2008-07-14 23:50:15 -06:00
|
|
|
case TPACKET_V2:
|
2010-11-30 19:52:20 -07:00
|
|
|
flush_dcache_page(pgv_to_page(&h.h2->tp_status));
|
2009-05-18 23:11:22 -06:00
|
|
|
return h.h2->tp_status;
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
2009-05-18 23:11:22 -06:00
|
|
|
default:
|
2011-08-19 04:18:16 -06:00
|
|
|
WARN(1, "TPACKET version not supported.\n");
|
2009-05-18 23:11:22 -06:00
|
|
|
BUG();
|
|
|
|
return 0;
|
2008-07-14 23:50:15 -06:00
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
static void *packet_lookup_frame(struct packet_sock *po,
|
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
unsigned int position,
|
|
|
|
int status)
|
|
|
|
{
|
|
|
|
unsigned int pg_vec_pos, frame_offset;
|
|
|
|
union {
|
|
|
|
struct tpacket_hdr *h1;
|
|
|
|
struct tpacket2_hdr *h2;
|
|
|
|
void *raw;
|
|
|
|
} h;
|
|
|
|
|
|
|
|
pg_vec_pos = position / rb->frames_per_block;
|
|
|
|
frame_offset = position % rb->frames_per_block;
|
|
|
|
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
h.raw = rb->pg_vec[pg_vec_pos].buffer +
|
|
|
|
(frame_offset * rb->frame_size);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
if (status != __packet_get_status(po, h.raw))
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
return h.raw;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *packet_current_frame(struct packet_sock *po,
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
int status)
|
|
|
|
{
|
|
|
|
return packet_lookup_frame(po, rb, rb->head, status);
|
|
|
|
}
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
del_timer_sync(&pkc->retire_blk_timer);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prb_shutdown_retire_blk_timer(struct packet_sock *po,
|
|
|
|
int tx_ring,
|
|
|
|
struct sk_buff_head *rb_queue)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
pkc = tx_ring ? &po->tx_ring.prb_bdqc : &po->rx_ring.prb_bdqc;
|
|
|
|
|
|
|
|
spin_lock(&rb_queue->lock);
|
|
|
|
pkc->delete_blk_timer = 1;
|
|
|
|
spin_unlock(&rb_queue->lock);
|
|
|
|
|
|
|
|
prb_del_retire_blk_timer(pkc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prb_init_blk_timer(struct packet_sock *po,
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
void (*func) (unsigned long))
|
|
|
|
{
|
|
|
|
init_timer(&pkc->retire_blk_timer);
|
|
|
|
pkc->retire_blk_timer.data = (long)po;
|
|
|
|
pkc->retire_blk_timer.function = func;
|
|
|
|
pkc->retire_blk_timer.expires = jiffies;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void prb_setup_retire_blk_timer(struct packet_sock *po, int tx_ring)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
if (tx_ring)
|
|
|
|
BUG();
|
|
|
|
|
|
|
|
pkc = tx_ring ? &po->tx_ring.prb_bdqc : &po->rx_ring.prb_bdqc;
|
|
|
|
prb_init_blk_timer(po, pkc, prb_retire_rx_blk_timer_expired);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int prb_calc_retire_blk_tmo(struct packet_sock *po,
|
|
|
|
int blk_size_in_bytes)
|
|
|
|
{
|
|
|
|
struct net_device *dev;
|
|
|
|
unsigned int mbits = 0, msec = 0, div = 0, tmo = 0;
|
2011-09-02 21:34:30 -06:00
|
|
|
struct ethtool_cmd ecmd;
|
|
|
|
int err;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
2011-09-02 21:34:30 -06:00
|
|
|
rtnl_lock();
|
|
|
|
dev = __dev_get_by_index(sock_net(&po->sk), po->ifindex);
|
|
|
|
if (unlikely(!dev)) {
|
|
|
|
rtnl_unlock();
|
2011-08-19 04:18:16 -06:00
|
|
|
return DEFAULT_PRB_RETIRE_TOV;
|
2011-09-02 21:34:30 -06:00
|
|
|
}
|
|
|
|
err = __ethtool_get_settings(dev, &ecmd);
|
|
|
|
rtnl_unlock();
|
|
|
|
if (!err) {
|
|
|
|
switch (ecmd.speed) {
|
|
|
|
case SPEED_10000:
|
|
|
|
msec = 1;
|
|
|
|
div = 10000/1000;
|
|
|
|
break;
|
|
|
|
case SPEED_1000:
|
|
|
|
msec = 1;
|
|
|
|
div = 1000/1000;
|
|
|
|
break;
|
|
|
|
/*
|
|
|
|
* If the link speed is so slow you don't really
|
|
|
|
* need to worry about perf anyways
|
|
|
|
*/
|
|
|
|
case SPEED_100:
|
|
|
|
case SPEED_10:
|
|
|
|
default:
|
|
|
|
return DEFAULT_PRB_RETIRE_TOV;
|
2011-08-19 04:18:16 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
mbits = (blk_size_in_bytes * 8) / (1024 * 1024);
|
|
|
|
|
|
|
|
if (div)
|
|
|
|
mbits /= div;
|
|
|
|
|
|
|
|
tmo = mbits * msec;
|
|
|
|
|
|
|
|
if (div)
|
|
|
|
return tmo+1;
|
|
|
|
return tmo;
|
|
|
|
}
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_init_ft_ops(struct tpacket_kbdq_core *p1,
|
2011-08-19 04:18:16 -06:00
|
|
|
union tpacket_req_u *req_u)
|
|
|
|
{
|
|
|
|
p1->feature_req_word = req_u->req3.tp_feature_req_word;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void init_prb_bdqc(struct packet_sock *po,
|
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
struct pgv *pg_vec,
|
|
|
|
union tpacket_req_u *req_u, int tx_ring)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *p1 = &rb->prb_bdqc;
|
|
|
|
struct tpacket_block_desc *pbd;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
memset(p1, 0x0, sizeof(*p1));
|
|
|
|
|
|
|
|
p1->knxt_seq_num = 1;
|
|
|
|
p1->pkbdq = pg_vec;
|
2011-08-25 04:43:30 -06:00
|
|
|
pbd = (struct tpacket_block_desc *)pg_vec[0].buffer;
|
2011-08-19 04:18:16 -06:00
|
|
|
p1->pkblk_start = (char *)pg_vec[0].buffer;
|
|
|
|
p1->kblk_size = req_u->req3.tp_block_size;
|
|
|
|
p1->knum_blocks = req_u->req3.tp_block_nr;
|
|
|
|
p1->hdrlen = po->tp_hdrlen;
|
|
|
|
p1->version = po->tp_version;
|
|
|
|
p1->last_kactive_blk_num = 0;
|
|
|
|
po->stats_u.stats3.tp_freeze_q_cnt = 0;
|
|
|
|
if (req_u->req3.tp_retire_blk_tov)
|
|
|
|
p1->retire_blk_tov = req_u->req3.tp_retire_blk_tov;
|
|
|
|
else
|
|
|
|
p1->retire_blk_tov = prb_calc_retire_blk_tmo(po,
|
|
|
|
req_u->req3.tp_block_size);
|
|
|
|
p1->tov_in_jiffies = msecs_to_jiffies(p1->retire_blk_tov);
|
|
|
|
p1->blk_sizeof_priv = req_u->req3.tp_sizeof_priv;
|
|
|
|
|
|
|
|
prb_init_ft_ops(p1, req_u);
|
|
|
|
prb_setup_retire_blk_timer(po, tx_ring);
|
|
|
|
prb_open_block(p1, pbd);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Do NOT update the last_blk_num first.
|
|
|
|
* Assumes sk_buff_head lock is held.
|
|
|
|
*/
|
2011-08-25 04:43:30 -06:00
|
|
|
static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
mod_timer(&pkc->retire_blk_timer,
|
|
|
|
jiffies + pkc->tov_in_jiffies);
|
|
|
|
pkc->last_kactive_blk_num = pkc->kactive_blk_num;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Timer logic:
|
|
|
|
* 1) We refresh the timer only when we open a block.
|
|
|
|
* By doing this we don't waste cycles refreshing the timer
|
|
|
|
* on packet-by-packet basis.
|
|
|
|
*
|
|
|
|
* With a 1MB block-size, on a 1Gbps line, it will take
|
|
|
|
* i) ~8 ms to fill a block + ii) memcpy etc.
|
|
|
|
* In this cut we are not accounting for the memcpy time.
|
|
|
|
*
|
|
|
|
* So, if the user sets the 'tmo' to 10ms then the timer
|
|
|
|
* will never fire while the block is still getting filled
|
|
|
|
* (which is what we want). However, the user could choose
|
|
|
|
* to close a block early and that's fine.
|
|
|
|
*
|
|
|
|
* But when the timer does fire, we check whether or not to refresh it.
|
|
|
|
* Since the tmo granularity is in msecs, it is not too expensive
|
|
|
|
* to refresh the timer, lets say every '8' msecs.
|
|
|
|
* Either the user can set the 'tmo' or we can derive it based on
|
|
|
|
* a) line-speed and b) block-size.
|
|
|
|
* prb_calc_retire_blk_tmo() calculates the tmo.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
static void prb_retire_rx_blk_timer_expired(unsigned long data)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = (struct packet_sock *)data;
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc = &po->rx_ring.prb_bdqc;
|
2011-08-19 04:18:16 -06:00
|
|
|
unsigned int frozen;
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_block_desc *pbd;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
spin_lock(&po->sk.sk_receive_queue.lock);
|
|
|
|
|
|
|
|
frozen = prb_queue_frozen(pkc);
|
|
|
|
pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
|
|
|
|
|
|
|
|
if (unlikely(pkc->delete_blk_timer))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/* We only need to plug the race when the block is partially filled.
|
|
|
|
* tpacket_rcv:
|
|
|
|
* lock(); increment BLOCK_NUM_PKTS; unlock()
|
|
|
|
* copy_bits() is in progress ...
|
|
|
|
* timer fires on other cpu:
|
|
|
|
* we can't retire the current block because copy_bits
|
|
|
|
* is in progress.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
if (BLOCK_NUM_PKTS(pbd)) {
|
|
|
|
while (atomic_read(&pkc->blk_fill_in_prog)) {
|
|
|
|
/* Waiting for skb_copy_bits to finish... */
|
|
|
|
cpu_relax();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (pkc->last_kactive_blk_num == pkc->kactive_blk_num) {
|
|
|
|
if (!frozen) {
|
|
|
|
prb_retire_current_block(pkc, po, TP_STATUS_BLK_TMO);
|
|
|
|
if (!prb_dispatch_next_block(pkc, po))
|
|
|
|
goto refresh_timer;
|
|
|
|
else
|
|
|
|
goto out;
|
|
|
|
} else {
|
|
|
|
/* Case 1. Queue was frozen because user-space was
|
|
|
|
* lagging behind.
|
|
|
|
*/
|
|
|
|
if (prb_curr_blk_in_use(pkc, pbd)) {
|
|
|
|
/*
|
|
|
|
* Ok, user-space is still behind.
|
|
|
|
* So just refresh the timer.
|
|
|
|
*/
|
|
|
|
goto refresh_timer;
|
|
|
|
} else {
|
|
|
|
/* Case 2. queue was frozen,user-space caught up,
|
|
|
|
* now the link went idle && the timer fired.
|
|
|
|
* We don't have a block to close.So we open this
|
|
|
|
* block and restart the timer.
|
|
|
|
* opening a block thaws the queue,restarts timer
|
|
|
|
* Thawing/timer-refresh is a side effect.
|
|
|
|
*/
|
|
|
|
prb_open_block(pkc, pbd);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
refresh_timer:
|
|
|
|
_prb_refresh_rx_retire_blk_timer(pkc);
|
|
|
|
|
|
|
|
out:
|
|
|
|
spin_unlock(&po->sk.sk_receive_queue.lock);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_flush_block(struct tpacket_kbdq_core *pkc1,
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_block_desc *pbd1, __u32 status)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
/* Flush everything minus the block header */
|
|
|
|
|
|
|
|
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
|
|
|
|
u8 *start, *end;
|
|
|
|
|
|
|
|
start = (u8 *)pbd1;
|
|
|
|
|
|
|
|
/* Skip the block header(we know header WILL fit in 4K) */
|
|
|
|
start += PAGE_SIZE;
|
|
|
|
|
|
|
|
end = (u8 *)PAGE_ALIGN((unsigned long)pkc1->pkblk_end);
|
|
|
|
for (; start < end; start += PAGE_SIZE)
|
|
|
|
flush_dcache_page(pgv_to_page(start));
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* Now update the block status. */
|
|
|
|
|
|
|
|
BLOCK_STATUS(pbd1) = status;
|
|
|
|
|
|
|
|
/* Flush the block header */
|
|
|
|
|
|
|
|
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
|
|
|
|
start = (u8 *)pbd1;
|
|
|
|
flush_dcache_page(pgv_to_page(start));
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Side effect:
|
|
|
|
*
|
|
|
|
* 1) flush the block
|
|
|
|
* 2) Increment active_blk_num
|
|
|
|
*
|
|
|
|
* Note:We DONT refresh the timer on purpose.
|
|
|
|
* Because almost always the next block will be opened.
|
|
|
|
*/
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_close_block(struct tpacket_kbdq_core *pkc1,
|
|
|
|
struct tpacket_block_desc *pbd1,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *po, unsigned int stat)
|
|
|
|
{
|
|
|
|
__u32 status = TP_STATUS_USER | stat;
|
|
|
|
|
|
|
|
struct tpacket3_hdr *last_pkt;
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
if (po->stats.tp_drops)
|
|
|
|
status |= TP_STATUS_LOSING;
|
|
|
|
|
|
|
|
last_pkt = (struct tpacket3_hdr *)pkc1->prev;
|
|
|
|
last_pkt->tp_next_offset = 0;
|
|
|
|
|
|
|
|
/* Get the ts of the last pkt */
|
|
|
|
if (BLOCK_NUM_PKTS(pbd1)) {
|
|
|
|
h1->ts_last_pkt.ts_sec = last_pkt->tp_sec;
|
|
|
|
h1->ts_last_pkt.ts_nsec = last_pkt->tp_nsec;
|
|
|
|
} else {
|
|
|
|
/* Ok, we tmo'd - so get the current time */
|
|
|
|
struct timespec ts;
|
|
|
|
getnstimeofday(&ts);
|
|
|
|
h1->ts_last_pkt.ts_sec = ts.tv_sec;
|
|
|
|
h1->ts_last_pkt.ts_nsec = ts.tv_nsec;
|
|
|
|
}
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
/* Flush the block */
|
|
|
|
prb_flush_block(pkc1, pbd1, status);
|
|
|
|
|
|
|
|
pkc1->kactive_blk_num = GET_NEXT_PRB_BLK_NUM(pkc1);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_thaw_queue(struct tpacket_kbdq_core *pkc)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
pkc->reset_pending_on_curr_blk = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Side effect of opening a block:
|
|
|
|
*
|
|
|
|
* 1) prb_queue is thawed.
|
|
|
|
* 2) retire_blk_timer is refreshed.
|
|
|
|
*
|
|
|
|
*/
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_open_block(struct tpacket_kbdq_core *pkc1,
|
|
|
|
struct tpacket_block_desc *pbd1)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
struct timespec ts;
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
smp_rmb();
|
|
|
|
|
|
|
|
if (likely(TP_STATUS_KERNEL == BLOCK_STATUS(pbd1))) {
|
|
|
|
|
|
|
|
/* We could have just memset this but we will lose the
|
|
|
|
* flexibility of making the priv area sticky
|
|
|
|
*/
|
|
|
|
BLOCK_SNUM(pbd1) = pkc1->knxt_seq_num++;
|
|
|
|
BLOCK_NUM_PKTS(pbd1) = 0;
|
|
|
|
BLOCK_LEN(pbd1) = BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
|
|
|
|
getnstimeofday(&ts);
|
|
|
|
h1->ts_first_pkt.ts_sec = ts.tv_sec;
|
|
|
|
h1->ts_first_pkt.ts_nsec = ts.tv_nsec;
|
|
|
|
pkc1->pkblk_start = (char *)pbd1;
|
|
|
|
pkc1->nxt_offset = (char *)(pkc1->pkblk_start +
|
|
|
|
BLK_PLUS_PRIV(pkc1->blk_sizeof_priv));
|
|
|
|
BLOCK_O2FP(pbd1) = (__u32)BLK_PLUS_PRIV(pkc1->blk_sizeof_priv);
|
|
|
|
BLOCK_O2PRIV(pbd1) = BLK_HDR_LEN;
|
|
|
|
pbd1->version = pkc1->version;
|
|
|
|
pkc1->prev = pkc1->nxt_offset;
|
|
|
|
pkc1->pkblk_end = pkc1->pkblk_start + pkc1->kblk_size;
|
|
|
|
prb_thaw_queue(pkc1);
|
|
|
|
_prb_refresh_rx_retire_blk_timer(pkc1);
|
|
|
|
|
|
|
|
smp_wmb();
|
|
|
|
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
WARN(1, "ERROR block:%p is NOT FREE status:%d kactive_blk_num:%d\n",
|
|
|
|
pbd1, BLOCK_STATUS(pbd1), pkc1->kactive_blk_num);
|
|
|
|
dump_stack();
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Queue freeze logic:
|
|
|
|
* 1) Assume tp_block_nr = 8 blocks.
|
|
|
|
* 2) At time 't0', user opens Rx ring.
|
|
|
|
* 3) Some time past 't0', kernel starts filling blocks starting from 0 .. 7
|
|
|
|
* 4) user-space is either sleeping or processing block '0'.
|
|
|
|
* 5) tpacket_rcv is currently filling block '7', since there is no space left,
|
|
|
|
* it will close block-7,loop around and try to fill block '0'.
|
|
|
|
* call-flow:
|
|
|
|
* __packet_lookup_frame_in_block
|
|
|
|
* prb_retire_current_block()
|
|
|
|
* prb_dispatch_next_block()
|
|
|
|
* |->(BLOCK_STATUS == USER) evaluates to true
|
|
|
|
* 5.1) Since block-0 is currently in-use, we just freeze the queue.
|
|
|
|
* 6) Now there are two cases:
|
|
|
|
* 6.1) Link goes idle right after the queue is frozen.
|
|
|
|
* But remember, the last open_block() refreshed the timer.
|
|
|
|
* When this timer expires,it will refresh itself so that we can
|
|
|
|
* re-open block-0 in near future.
|
|
|
|
* 6.2) Link is busy and keeps on receiving packets. This is a simple
|
|
|
|
* case and __packet_lookup_frame_in_block will check if block-0
|
|
|
|
* is free and can now be re-used.
|
|
|
|
*/
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_freeze_queue(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *po)
|
|
|
|
{
|
|
|
|
pkc->reset_pending_on_curr_blk = 1;
|
|
|
|
po->stats_u.stats3.tp_freeze_q_cnt++;
|
|
|
|
}
|
|
|
|
|
|
|
|
#define TOTAL_PKT_LEN_INCL_ALIGN(length) (ALIGN((length), V3_ALIGNMENT))
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the next block is free then we will dispatch it
|
|
|
|
* and return a good offset.
|
|
|
|
* Else, we will freeze the queue.
|
|
|
|
* So, caller must check the return value.
|
|
|
|
*/
|
2011-08-25 04:43:30 -06:00
|
|
|
static void *prb_dispatch_next_block(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *po)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_block_desc *pbd;
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
smp_rmb();
|
|
|
|
|
|
|
|
/* 1. Get current block num */
|
|
|
|
pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
|
|
|
|
|
|
|
|
/* 2. If this block is currently in_use then freeze the queue */
|
|
|
|
if (TP_STATUS_USER & BLOCK_STATUS(pbd)) {
|
|
|
|
prb_freeze_queue(pkc, po);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 3.
|
|
|
|
* open this block and return the offset where the first packet
|
|
|
|
* needs to get stored.
|
|
|
|
*/
|
|
|
|
prb_open_block(pkc, pbd);
|
|
|
|
return (void *)pkc->nxt_offset;
|
|
|
|
}
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_retire_current_block(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_sock *po, unsigned int status)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_block_desc *pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
/* retire/close the current block */
|
|
|
|
if (likely(TP_STATUS_KERNEL == BLOCK_STATUS(pbd))) {
|
|
|
|
/*
|
|
|
|
* Plug the case where copy_bits() is in progress on
|
|
|
|
* cpu-0 and tpacket_rcv() got invoked on cpu-1, didn't
|
|
|
|
* have space to copy the pkt in the current block and
|
|
|
|
* called prb_retire_current_block()
|
|
|
|
*
|
|
|
|
* We don't need to worry about the TMO case because
|
|
|
|
* the timer-handler already handled this case.
|
|
|
|
*/
|
|
|
|
if (!(status & TP_STATUS_BLK_TMO)) {
|
|
|
|
while (atomic_read(&pkc->blk_fill_in_prog)) {
|
|
|
|
/* Waiting for skb_copy_bits to finish... */
|
|
|
|
cpu_relax();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
prb_close_block(pkc, pbd, po, status);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
WARN(1, "ERROR-pbd[%d]:%p\n", pkc->kactive_blk_num, pbd);
|
|
|
|
dump_stack();
|
|
|
|
BUG();
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static int prb_curr_blk_in_use(struct tpacket_kbdq_core *pkc,
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_block_desc *pbd)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
return TP_STATUS_USER & BLOCK_STATUS(pbd);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static int prb_queue_frozen(struct tpacket_kbdq_core *pkc)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
return pkc->reset_pending_on_curr_blk;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_clear_blk_fill_status(struct packet_ring_buffer *rb)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc = GET_PBDQC_FROM_RB(rb);
|
2011-08-19 04:18:16 -06:00
|
|
|
atomic_dec(&pkc->blk_fill_in_prog);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_fill_rxhash(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct tpacket3_hdr *ppd)
|
|
|
|
{
|
|
|
|
ppd->hv1.tp_rxhash = skb_get_rxhash(pkc->skb);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_clear_rxhash(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct tpacket3_hdr *ppd)
|
|
|
|
{
|
|
|
|
ppd->hv1.tp_rxhash = 0;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_fill_vlan_info(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct tpacket3_hdr *ppd)
|
|
|
|
{
|
|
|
|
if (vlan_tx_tag_present(pkc->skb)) {
|
|
|
|
ppd->hv1.tp_vlan_tci = vlan_tx_tag_get(pkc->skb);
|
|
|
|
ppd->tp_status = TP_STATUS_VLAN_VALID;
|
|
|
|
} else {
|
|
|
|
ppd->hv1.tp_vlan_tci = ppd->tp_status = 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-08-25 04:43:30 -06:00
|
|
|
static void prb_run_all_ft_ops(struct tpacket_kbdq_core *pkc,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct tpacket3_hdr *ppd)
|
|
|
|
{
|
|
|
|
prb_fill_vlan_info(pkc, ppd);
|
|
|
|
|
|
|
|
if (pkc->feature_req_word & TP_FT_REQ_FILL_RXHASH)
|
|
|
|
prb_fill_rxhash(pkc, ppd);
|
|
|
|
else
|
|
|
|
prb_clear_rxhash(pkc, ppd);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void prb_fill_curr_block(char *curr,
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc,
|
|
|
|
struct tpacket_block_desc *pbd,
|
2011-08-19 04:18:16 -06:00
|
|
|
unsigned int len)
|
|
|
|
{
|
|
|
|
struct tpacket3_hdr *ppd;
|
|
|
|
|
|
|
|
ppd = (struct tpacket3_hdr *)curr;
|
|
|
|
ppd->tp_next_offset = TOTAL_PKT_LEN_INCL_ALIGN(len);
|
|
|
|
pkc->prev = curr;
|
|
|
|
pkc->nxt_offset += TOTAL_PKT_LEN_INCL_ALIGN(len);
|
|
|
|
BLOCK_LEN(pbd) += TOTAL_PKT_LEN_INCL_ALIGN(len);
|
|
|
|
BLOCK_NUM_PKTS(pbd) += 1;
|
|
|
|
atomic_inc(&pkc->blk_fill_in_prog);
|
|
|
|
prb_run_all_ft_ops(pkc, ppd);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Assumes caller has the sk->rx_queue.lock */
|
|
|
|
static void *__packet_lookup_frame_in_block(struct packet_sock *po,
|
|
|
|
struct sk_buff *skb,
|
|
|
|
int status,
|
|
|
|
unsigned int len
|
|
|
|
)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc;
|
|
|
|
struct tpacket_block_desc *pbd;
|
2011-08-19 04:18:16 -06:00
|
|
|
char *curr, *end;
|
|
|
|
|
|
|
|
pkc = GET_PBDQC_FROM_RB(((struct packet_ring_buffer *)&po->rx_ring));
|
|
|
|
pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
|
|
|
|
|
|
|
|
/* Queue is frozen when user space is lagging behind */
|
|
|
|
if (prb_queue_frozen(pkc)) {
|
|
|
|
/*
|
|
|
|
* Check if that last block which caused the queue to freeze,
|
|
|
|
* is still in_use by user-space.
|
|
|
|
*/
|
|
|
|
if (prb_curr_blk_in_use(pkc, pbd)) {
|
|
|
|
/* Can't record this packet */
|
|
|
|
return NULL;
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Ok, the block was released by user-space.
|
|
|
|
* Now let's open that block.
|
|
|
|
* opening a block also thaws the queue.
|
|
|
|
* Thawing is a side effect.
|
|
|
|
*/
|
|
|
|
prb_open_block(pkc, pbd);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
smp_mb();
|
|
|
|
curr = pkc->nxt_offset;
|
|
|
|
pkc->skb = skb;
|
|
|
|
end = (char *) ((char *)pbd + pkc->kblk_size);
|
|
|
|
|
|
|
|
/* first try the current block */
|
|
|
|
if (curr+TOTAL_PKT_LEN_INCL_ALIGN(len) < end) {
|
|
|
|
prb_fill_curr_block(curr, pkc, pbd, len);
|
|
|
|
return (void *)curr;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Ok, close the current block */
|
|
|
|
prb_retire_current_block(pkc, po, 0);
|
|
|
|
|
|
|
|
/* Now, try to dispatch the next block */
|
|
|
|
curr = (char *)prb_dispatch_next_block(pkc, po);
|
|
|
|
if (curr) {
|
|
|
|
pbd = GET_CURR_PBLOCK_DESC_FROM_CORE(pkc);
|
|
|
|
prb_fill_curr_block(curr, pkc, pbd, len);
|
|
|
|
return (void *)curr;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* No free blocks are available.user_space hasn't caught up yet.
|
|
|
|
* Queue was just frozen and now this packet will get dropped.
|
|
|
|
*/
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *packet_current_rx_frame(struct packet_sock *po,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct sk_buff *skb,
|
|
|
|
int status, unsigned int len)
|
|
|
|
{
|
|
|
|
char *curr = NULL;
|
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
case TPACKET_V2:
|
|
|
|
curr = packet_lookup_frame(po, &po->rx_ring,
|
|
|
|
po->rx_ring.head, status);
|
|
|
|
return curr;
|
|
|
|
case TPACKET_V3:
|
|
|
|
return __packet_lookup_frame_in_block(po, skb, status, len);
|
|
|
|
default:
|
|
|
|
WARN(1, "TPACKET version not supported\n");
|
|
|
|
BUG();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *prb_lookup_block(struct packet_sock *po,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
unsigned int previous,
|
|
|
|
int status)
|
|
|
|
{
|
2011-08-25 04:43:30 -06:00
|
|
|
struct tpacket_kbdq_core *pkc = GET_PBDQC_FROM_RB(rb);
|
|
|
|
struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, previous);
|
2011-08-19 04:18:16 -06:00
|
|
|
|
|
|
|
if (status != BLOCK_STATUS(pbd))
|
|
|
|
return NULL;
|
|
|
|
return pbd;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static int prb_previous_blk_num(struct packet_ring_buffer *rb)
|
2011-08-19 04:18:16 -06:00
|
|
|
{
|
|
|
|
unsigned int prev;
|
|
|
|
if (rb->prb_bdqc.kactive_blk_num)
|
|
|
|
prev = rb->prb_bdqc.kactive_blk_num-1;
|
|
|
|
else
|
|
|
|
prev = rb->prb_bdqc.knum_blocks-1;
|
|
|
|
return prev;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Assumes caller has held the rx_queue.lock */
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *__prb_previous_block(struct packet_sock *po,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
int status)
|
|
|
|
{
|
|
|
|
unsigned int previous = prb_previous_blk_num(rb);
|
|
|
|
return prb_lookup_block(po, rb, previous, status);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *packet_previous_rx_frame(struct packet_sock *po,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
int status)
|
|
|
|
{
|
|
|
|
if (po->tp_version <= TPACKET_V2)
|
|
|
|
return packet_previous_frame(po, rb, status);
|
|
|
|
|
|
|
|
return __prb_previous_block(po, rb, status);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void packet_increment_rx_head(struct packet_sock *po,
|
2011-08-19 04:18:16 -06:00
|
|
|
struct packet_ring_buffer *rb)
|
|
|
|
{
|
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
case TPACKET_V2:
|
|
|
|
return packet_increment_head(rb);
|
|
|
|
case TPACKET_V3:
|
|
|
|
default:
|
|
|
|
WARN(1, "TPACKET version not supported.\n");
|
|
|
|
BUG();
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void *packet_previous_frame(struct packet_sock *po,
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_ring_buffer *rb,
|
|
|
|
int status)
|
|
|
|
{
|
|
|
|
unsigned int previous = rb->head ? rb->head - 1 : rb->frame_max;
|
|
|
|
return packet_lookup_frame(po, rb, previous, status);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static void packet_increment_head(struct packet_ring_buffer *buff)
|
2009-05-18 23:11:22 -06:00
|
|
|
{
|
|
|
|
buff->head = buff->head != buff->frame_max ? buff->head+1 : 0;
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
static void packet_sock_destruct(struct sock *sk)
|
|
|
|
{
|
2010-04-07 16:41:28 -06:00
|
|
|
skb_queue_purge(&sk->sk_error_queue);
|
|
|
|
|
2008-07-25 22:43:18 -06:00
|
|
|
WARN_ON(atomic_read(&sk->sk_rmem_alloc));
|
|
|
|
WARN_ON(atomic_read(&sk->sk_wmem_alloc));
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (!sock_flag(sk, SOCK_DEAD)) {
|
2009-07-21 15:57:59 -06:00
|
|
|
pr_err("Attempt to release alive packet socket: %p\n", sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2007-11-10 22:38:48 -07:00
|
|
|
sk_refcnt_debug_dec(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
static int fanout_rr_next(struct packet_fanout *f, unsigned int num)
|
|
|
|
{
|
|
|
|
int x = atomic_read(&f->rr_cur) + 1;
|
|
|
|
|
|
|
|
if (x >= num)
|
|
|
|
x = 0;
|
|
|
|
|
|
|
|
return x;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct sock *fanout_demux_hash(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
|
|
|
|
{
|
|
|
|
u32 idx, hash = skb->rxhash;
|
|
|
|
|
|
|
|
idx = ((u64)hash * num) >> 32;
|
|
|
|
|
|
|
|
return f->arr[idx];
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct sock *fanout_demux_lb(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
|
|
|
|
{
|
|
|
|
int cur, old;
|
|
|
|
|
|
|
|
cur = atomic_read(&f->rr_cur);
|
|
|
|
while ((old = atomic_cmpxchg(&f->rr_cur, cur,
|
|
|
|
fanout_rr_next(f, num))) != cur)
|
|
|
|
cur = old;
|
|
|
|
return f->arr[cur];
|
|
|
|
}
|
|
|
|
|
2011-07-06 02:56:38 -06:00
|
|
|
static struct sock *fanout_demux_cpu(struct packet_fanout *f, struct sk_buff *skb, unsigned int num)
|
|
|
|
{
|
|
|
|
unsigned int cpu = smp_processor_id();
|
|
|
|
|
|
|
|
return f->arr[cpu % num];
|
|
|
|
}
|
|
|
|
|
|
|
|
static int packet_rcv_fanout(struct sk_buff *skb, struct net_device *dev,
|
|
|
|
struct packet_type *pt, struct net_device *orig_dev)
|
2011-07-05 02:45:05 -06:00
|
|
|
{
|
|
|
|
struct packet_fanout *f = pt->af_packet_priv;
|
|
|
|
unsigned int num = f->num_members;
|
|
|
|
struct packet_sock *po;
|
|
|
|
struct sock *sk;
|
|
|
|
|
|
|
|
if (!net_eq(dev_net(dev), read_pnet(&f->net)) ||
|
|
|
|
!num) {
|
|
|
|
kfree_skb(skb);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-07-06 02:56:38 -06:00
|
|
|
switch (f->type) {
|
|
|
|
case PACKET_FANOUT_HASH:
|
|
|
|
default:
|
|
|
|
if (f->defrag) {
|
2011-10-06 04:28:31 -06:00
|
|
|
skb = ip_check_defrag(skb, IP_DEFRAG_AF_PACKET);
|
2011-07-06 02:56:38 -06:00
|
|
|
if (!skb)
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
skb_get_rxhash(skb);
|
|
|
|
sk = fanout_demux_hash(f, skb, num);
|
|
|
|
break;
|
|
|
|
case PACKET_FANOUT_LB:
|
|
|
|
sk = fanout_demux_lb(f, skb, num);
|
|
|
|
break;
|
|
|
|
case PACKET_FANOUT_CPU:
|
|
|
|
sk = fanout_demux_cpu(f, skb, num);
|
|
|
|
break;
|
2011-07-05 02:45:05 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
po = pkt_sk(sk);
|
|
|
|
|
|
|
|
return po->prot_hook.func(skb, dev, &po->prot_hook, orig_dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static DEFINE_MUTEX(fanout_mutex);
|
|
|
|
static LIST_HEAD(fanout_list);
|
|
|
|
|
|
|
|
static void __fanout_link(struct sock *sk, struct packet_sock *po)
|
|
|
|
{
|
|
|
|
struct packet_fanout *f = po->fanout;
|
|
|
|
|
|
|
|
spin_lock(&f->lock);
|
|
|
|
f->arr[f->num_members] = sk;
|
|
|
|
smp_wmb();
|
|
|
|
f->num_members++;
|
|
|
|
spin_unlock(&f->lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void __fanout_unlink(struct sock *sk, struct packet_sock *po)
|
|
|
|
{
|
|
|
|
struct packet_fanout *f = po->fanout;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
spin_lock(&f->lock);
|
|
|
|
for (i = 0; i < f->num_members; i++) {
|
|
|
|
if (f->arr[i] == sk)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
BUG_ON(i >= f->num_members);
|
|
|
|
f->arr[i] = f->arr[f->num_members - 1];
|
|
|
|
f->num_members--;
|
|
|
|
spin_unlock(&f->lock);
|
|
|
|
}
|
|
|
|
|
2011-07-05 02:43:20 -06:00
|
|
|
static int fanout_add(struct sock *sk, u16 id, u16 type_flags)
|
2011-07-05 02:45:05 -06:00
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
struct packet_fanout *f, *match;
|
2011-07-05 02:43:20 -06:00
|
|
|
u8 type = type_flags & 0xff;
|
|
|
|
u8 defrag = (type_flags & PACKET_FANOUT_FLAG_DEFRAG) ? 1 : 0;
|
2011-07-05 02:45:05 -06:00
|
|
|
int err;
|
|
|
|
|
|
|
|
switch (type) {
|
|
|
|
case PACKET_FANOUT_HASH:
|
|
|
|
case PACKET_FANOUT_LB:
|
2011-07-06 02:56:38 -06:00
|
|
|
case PACKET_FANOUT_CPU:
|
2011-07-05 02:45:05 -06:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!po->running)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (po->fanout)
|
|
|
|
return -EALREADY;
|
|
|
|
|
|
|
|
mutex_lock(&fanout_mutex);
|
|
|
|
match = NULL;
|
|
|
|
list_for_each_entry(f, &fanout_list, list) {
|
|
|
|
if (f->id == id &&
|
|
|
|
read_pnet(&f->net) == sock_net(sk)) {
|
|
|
|
match = f;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2011-07-07 07:41:29 -06:00
|
|
|
err = -EINVAL;
|
2011-07-05 02:43:20 -06:00
|
|
|
if (match && match->defrag != defrag)
|
2011-07-07 07:41:29 -06:00
|
|
|
goto out;
|
2011-07-05 02:45:05 -06:00
|
|
|
if (!match) {
|
2011-07-07 07:41:29 -06:00
|
|
|
err = -ENOMEM;
|
2011-07-05 02:45:05 -06:00
|
|
|
match = kzalloc(sizeof(*match), GFP_KERNEL);
|
2011-07-07 07:41:29 -06:00
|
|
|
if (!match)
|
|
|
|
goto out;
|
|
|
|
write_pnet(&match->net, sock_net(sk));
|
|
|
|
match->id = id;
|
|
|
|
match->type = type;
|
|
|
|
match->defrag = defrag;
|
|
|
|
atomic_set(&match->rr_cur, 0);
|
|
|
|
INIT_LIST_HEAD(&match->list);
|
|
|
|
spin_lock_init(&match->lock);
|
|
|
|
atomic_set(&match->sk_ref, 0);
|
|
|
|
match->prot_hook.type = po->prot_hook.type;
|
|
|
|
match->prot_hook.dev = po->prot_hook.dev;
|
|
|
|
match->prot_hook.func = packet_rcv_fanout;
|
|
|
|
match->prot_hook.af_packet_priv = match;
|
|
|
|
dev_add_pack(&match->prot_hook);
|
|
|
|
list_add(&match->list, &fanout_list);
|
2011-07-05 02:45:05 -06:00
|
|
|
}
|
2011-07-07 07:41:29 -06:00
|
|
|
err = -EINVAL;
|
|
|
|
if (match->type == type &&
|
|
|
|
match->prot_hook.type == po->prot_hook.type &&
|
|
|
|
match->prot_hook.dev == po->prot_hook.dev) {
|
|
|
|
err = -ENOSPC;
|
|
|
|
if (atomic_read(&match->sk_ref) < PACKET_FANOUT_MAX) {
|
|
|
|
__dev_remove_pack(&po->prot_hook);
|
|
|
|
po->fanout = match;
|
|
|
|
atomic_inc(&match->sk_ref);
|
|
|
|
__fanout_link(sk, po);
|
|
|
|
err = 0;
|
2011-07-05 02:45:05 -06:00
|
|
|
}
|
|
|
|
}
|
2011-07-07 07:41:29 -06:00
|
|
|
out:
|
2011-07-05 02:45:05 -06:00
|
|
|
mutex_unlock(&fanout_mutex);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void fanout_release(struct sock *sk)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
struct packet_fanout *f;
|
|
|
|
|
|
|
|
f = po->fanout;
|
|
|
|
if (!f)
|
|
|
|
return;
|
|
|
|
|
|
|
|
po->fanout = NULL;
|
|
|
|
|
|
|
|
mutex_lock(&fanout_mutex);
|
|
|
|
if (atomic_dec_and_test(&f->sk_ref)) {
|
|
|
|
list_del(&f->list);
|
|
|
|
dev_remove_pack(&f->prot_hook);
|
|
|
|
kfree(f);
|
|
|
|
}
|
|
|
|
mutex_unlock(&fanout_mutex);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-12-22 13:49:22 -07:00
|
|
|
static const struct proto_ops packet_ops;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-12-22 13:49:22 -07:00
|
|
|
static const struct proto_ops packet_ops_spkt;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static int packet_rcv_spkt(struct sk_buff *skb, struct net_device *dev,
|
|
|
|
struct packet_type *pt, struct net_device *orig_dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk;
|
|
|
|
struct sockaddr_pkt *spkt;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When we registered the protocol we saved the socket in the data
|
|
|
|
* field for just this event.
|
|
|
|
*/
|
|
|
|
|
|
|
|
sk = pt->af_packet_priv;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Yank back the headers [hope the device set this
|
|
|
|
* right or kerboom...]
|
|
|
|
*
|
|
|
|
* Incoming packets have ll header pulled,
|
|
|
|
* push it back.
|
|
|
|
*
|
2007-03-19 16:33:04 -06:00
|
|
|
* For outgoing ones skb->data == skb_mac_header(skb)
|
2005-04-16 16:20:36 -06:00
|
|
|
* so that this procedure is noop.
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (skb->pkt_type == PACKET_LOOPBACK)
|
|
|
|
goto out;
|
|
|
|
|
2009-11-25 16:14:13 -07:00
|
|
|
if (!net_eq(dev_net(dev), sock_net(sk)))
|
2007-11-19 23:28:35 -07:00
|
|
|
goto out;
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
skb = skb_share_check(skb, GFP_ATOMIC);
|
|
|
|
if (skb == NULL)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto oom;
|
|
|
|
|
|
|
|
/* drop any routing info */
|
2009-06-01 23:19:30 -06:00
|
|
|
skb_dst_drop(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-07-12 12:57:52 -06:00
|
|
|
/* drop conntrack reference */
|
|
|
|
nf_reset(skb);
|
|
|
|
|
2007-02-05 00:33:10 -07:00
|
|
|
spkt = &PACKET_SKB_CB(skb)->sa.pkt;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-03-19 16:33:04 -06:00
|
|
|
skb_push(skb, skb->data - skb_mac_header(skb));
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/*
|
|
|
|
* The SOCK_PACKET socket receives _all_ frames.
|
|
|
|
*/
|
|
|
|
|
|
|
|
spkt->spkt_family = dev->type;
|
|
|
|
strlcpy(spkt->spkt_device, dev->name, sizeof(spkt->spkt_device));
|
|
|
|
spkt->spkt_protocol = skb->protocol;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Charge the memory to the socket. This is done specifically
|
|
|
|
* to prevent sockets using all the memory up.
|
|
|
|
*/
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
if (sock_queue_rcv_skb(sk, skb) == 0)
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
out:
|
|
|
|
kfree_skb(skb);
|
|
|
|
oom:
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Output a raw packet to a device layer. This bypasses all the other
|
|
|
|
* protocol layers and you must therefore supply it with a complete frame
|
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
static int packet_sendmsg_spkt(struct kiocb *iocb, struct socket *sock,
|
|
|
|
struct msghdr *msg, size_t len)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
2009-07-21 15:57:59 -06:00
|
|
|
struct sockaddr_pkt *saddr = (struct sockaddr_pkt *)msg->msg_name;
|
2009-12-14 22:47:03 -07:00
|
|
|
struct sk_buff *skb = NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct net_device *dev;
|
2009-07-21 15:57:59 -06:00
|
|
|
__be16 proto = 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
int err;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
2007-02-09 07:25:10 -07:00
|
|
|
* Get and verify the address.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
if (saddr) {
|
2005-04-16 16:20:36 -06:00
|
|
|
if (msg->msg_namelen < sizeof(struct sockaddr))
|
2009-07-21 15:57:59 -06:00
|
|
|
return -EINVAL;
|
|
|
|
if (msg->msg_namelen == sizeof(struct sockaddr_pkt))
|
|
|
|
proto = saddr->spkt_protocol;
|
|
|
|
} else
|
|
|
|
return -ENOTCONN; /* SOCK_PACKET must be sent giving an address */
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/*
|
2007-02-09 07:25:10 -07:00
|
|
|
* Find the device first to size check it
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
|
|
|
|
saddr->spkt_device[13] = 0;
|
2009-12-14 22:47:03 -07:00
|
|
|
retry:
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
dev = dev_get_by_name_rcu(sock_net(sk), saddr->spkt_device);
|
2005-04-16 16:20:36 -06:00
|
|
|
err = -ENODEV;
|
|
|
|
if (dev == NULL)
|
|
|
|
goto out_unlock;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2007-01-25 20:30:36 -07:00
|
|
|
err = -ENETDOWN;
|
|
|
|
if (!(dev->flags & IFF_UP))
|
|
|
|
goto out_unlock;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
2009-07-21 15:57:59 -06:00
|
|
|
* You may not queue a frame bigger than the mtu. This is the lowest level
|
|
|
|
* raw protocol and you must do your own fragmentation at this level.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
err = -EMSGSIZE;
|
2011-02-11 02:35:18 -07:00
|
|
|
if (len > dev->mtu + dev->hard_header_len + VLAN_HLEN)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out_unlock;
|
|
|
|
|
2009-12-14 22:47:03 -07:00
|
|
|
if (!skb) {
|
|
|
|
size_t reserved = LL_RESERVED_SPACE(dev);
|
2011-11-17 19:20:05 -07:00
|
|
|
int tlen = dev->needed_tailroom;
|
2009-12-14 22:47:03 -07:00
|
|
|
unsigned int hhlen = dev->header_ops ? dev->hard_header_len : 0;
|
|
|
|
|
|
|
|
rcu_read_unlock();
|
2011-11-17 19:20:05 -07:00
|
|
|
skb = sock_wmalloc(sk, len + reserved + tlen, 0, GFP_KERNEL);
|
2009-12-14 22:47:03 -07:00
|
|
|
if (skb == NULL)
|
|
|
|
return -ENOBUFS;
|
|
|
|
/* FIXME: Save some space for broken drivers that write a hard
|
|
|
|
* header at transmission time by themselves. PPP is the notable
|
|
|
|
* one here. This should really be fixed at the driver level.
|
|
|
|
*/
|
|
|
|
skb_reserve(skb, reserved);
|
|
|
|
skb_reset_network_header(skb);
|
|
|
|
|
|
|
|
/* Try to align data part correctly */
|
|
|
|
if (hhlen) {
|
|
|
|
skb->data -= hhlen;
|
|
|
|
skb->tail -= hhlen;
|
|
|
|
if (len < hhlen)
|
|
|
|
skb_reset_network_header(skb);
|
|
|
|
}
|
|
|
|
err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
|
|
|
|
if (err)
|
|
|
|
goto out_free;
|
|
|
|
goto retry;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2011-02-11 02:35:18 -07:00
|
|
|
if (len > (dev->mtu + dev->hard_header_len)) {
|
|
|
|
/* Earlier code assumed this would be a VLAN pkt,
|
|
|
|
* double-check this now that we have the actual
|
|
|
|
* packet in hand.
|
|
|
|
*/
|
|
|
|
struct ethhdr *ehdr;
|
|
|
|
skb_reset_mac_header(skb);
|
|
|
|
ehdr = eth_hdr(skb);
|
|
|
|
if (ehdr->h_proto != htons(ETH_P_8021Q)) {
|
|
|
|
err = -EMSGSIZE;
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
|
|
|
}
|
2009-12-14 22:47:03 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
skb->protocol = proto;
|
|
|
|
skb->dev = dev;
|
|
|
|
skb->priority = sk->sk_priority;
|
2009-10-01 13:14:46 -06:00
|
|
|
skb->mark = sk->sk_mark;
|
2010-08-17 02:59:14 -06:00
|
|
|
err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
|
2010-04-07 16:41:28 -06:00
|
|
|
if (err < 0)
|
|
|
|
goto out_unlock;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
dev_queue_xmit(skb);
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_unlock();
|
2009-07-21 15:57:59 -06:00
|
|
|
return len;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
out_unlock:
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_unlock();
|
2009-12-14 22:47:03 -07:00
|
|
|
out_free:
|
|
|
|
kfree_skb(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static unsigned int run_filter(const struct sk_buff *skb,
|
2010-12-06 13:50:09 -07:00
|
|
|
const struct sock *sk,
|
2007-01-24 16:21:02 -07:00
|
|
|
unsigned int res)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sk_filter *filter;
|
2006-08-31 16:28:39 -06:00
|
|
|
|
2011-01-18 00:46:52 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
filter = rcu_dereference(sk->sk_filter);
|
2007-01-24 16:21:02 -07:00
|
|
|
if (filter != NULL)
|
2011-04-20 03:27:32 -06:00
|
|
|
res = SK_RUN_FILTER(filter, skb);
|
2011-01-18 00:46:52 -07:00
|
|
|
rcu_read_unlock();
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-01-24 16:21:02 -07:00
|
|
|
return res;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2010-12-06 13:50:09 -07:00
|
|
|
* This function makes lazy skb cloning in hope that most of packets
|
|
|
|
* are discarded by BPF.
|
|
|
|
*
|
|
|
|
* Note tricky part: we DO mangle shared skb! skb->data, skb->len
|
|
|
|
* and skb->cb are mangled. It works because (and until) packets
|
|
|
|
* falling here are owned by current CPU. Output packets are cloned
|
|
|
|
* by dev_queue_xmit_nit(), input packets are processed by net_bh
|
|
|
|
* sequencially, so that if we return skb to original state on exit,
|
|
|
|
* we will not harm anyone.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static int packet_rcv(struct sk_buff *skb, struct net_device *dev,
|
|
|
|
struct packet_type *pt, struct net_device *orig_dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk;
|
|
|
|
struct sockaddr_ll *sll;
|
|
|
|
struct packet_sock *po;
|
2009-07-21 15:57:59 -06:00
|
|
|
u8 *skb_head = skb->data;
|
2005-04-16 16:20:36 -06:00
|
|
|
int skb_len = skb->len;
|
2007-01-24 16:21:02 -07:00
|
|
|
unsigned int snaplen, res;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (skb->pkt_type == PACKET_LOOPBACK)
|
|
|
|
goto drop;
|
|
|
|
|
|
|
|
sk = pt->af_packet_priv;
|
|
|
|
po = pkt_sk(sk);
|
|
|
|
|
2009-11-25 16:14:13 -07:00
|
|
|
if (!net_eq(dev_net(dev), sock_net(sk)))
|
2007-11-19 23:28:35 -07:00
|
|
|
goto drop;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
skb->dev = dev;
|
|
|
|
|
2007-10-09 02:40:57 -06:00
|
|
|
if (dev->header_ops) {
|
2005-04-16 16:20:36 -06:00
|
|
|
/* The device has an explicit notion of ll header,
|
2010-12-06 13:50:09 -07:00
|
|
|
* exported to higher levels.
|
|
|
|
*
|
|
|
|
* Otherwise, the device hides details of its frame
|
|
|
|
* structure, so that corresponding packet head is
|
|
|
|
* never delivered to user.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
if (sk->sk_type != SOCK_DGRAM)
|
2007-03-19 16:33:04 -06:00
|
|
|
skb_push(skb, skb->data - skb_mac_header(skb));
|
2005-04-16 16:20:36 -06:00
|
|
|
else if (skb->pkt_type == PACKET_OUTGOING) {
|
|
|
|
/* Special case: outgoing packets have ll header at head */
|
2007-03-10 18:16:10 -07:00
|
|
|
skb_pull(skb, skb_network_offset(skb));
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
snaplen = skb->len;
|
|
|
|
|
2007-01-24 16:21:02 -07:00
|
|
|
res = run_filter(skb, sk, snaplen);
|
|
|
|
if (!res)
|
2006-08-31 16:28:39 -06:00
|
|
|
goto drop_n_restore;
|
2007-01-24 16:21:02 -07:00
|
|
|
if (snaplen > res)
|
|
|
|
snaplen = res;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-12-21 00:11:44 -07:00
|
|
|
if (atomic_read(&sk->sk_rmem_alloc) >= sk->sk_rcvbuf)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto drop_n_acct;
|
|
|
|
|
|
|
|
if (skb_shared(skb)) {
|
|
|
|
struct sk_buff *nskb = skb_clone(skb, GFP_ATOMIC);
|
|
|
|
if (nskb == NULL)
|
|
|
|
goto drop_n_acct;
|
|
|
|
|
|
|
|
if (skb_head != skb->data) {
|
|
|
|
skb->data = skb_head;
|
|
|
|
skb->len = skb_len;
|
|
|
|
}
|
|
|
|
kfree_skb(skb);
|
|
|
|
skb = nskb;
|
|
|
|
}
|
|
|
|
|
2007-02-05 00:33:10 -07:00
|
|
|
BUILD_BUG_ON(sizeof(*PACKET_SKB_CB(skb)) + MAX_ADDR_LEN - 8 >
|
|
|
|
sizeof(skb->cb));
|
|
|
|
|
|
|
|
sll = &PACKET_SKB_CB(skb)->sa.ll;
|
2005-04-16 16:20:36 -06:00
|
|
|
sll->sll_family = AF_PACKET;
|
|
|
|
sll->sll_hatype = dev->type;
|
|
|
|
sll->sll_protocol = skb->protocol;
|
|
|
|
sll->sll_pkttype = skb->pkt_type;
|
2007-11-10 23:03:25 -07:00
|
|
|
if (unlikely(po->origdev))
|
2007-04-20 17:05:39 -06:00
|
|
|
sll->sll_ifindex = orig_dev->ifindex;
|
|
|
|
else
|
|
|
|
sll->sll_ifindex = dev->ifindex;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-09-26 23:13:38 -06:00
|
|
|
sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-02-05 00:33:10 -07:00
|
|
|
PACKET_SKB_CB(skb)->origlen = skb->len;
|
2007-02-05 00:31:32 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (pskb_trim(skb, snaplen))
|
|
|
|
goto drop_n_acct;
|
|
|
|
|
|
|
|
skb_set_owner_r(skb, sk);
|
|
|
|
skb->dev = NULL;
|
2009-06-01 23:19:30 -06:00
|
|
|
skb_dst_drop(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-07-12 12:57:52 -06:00
|
|
|
/* drop conntrack reference */
|
|
|
|
nf_reset(skb);
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
spin_lock(&sk->sk_receive_queue.lock);
|
|
|
|
po->stats.tp_packets++;
|
net: Generalize socket rx gap / receive queue overflow cmsg
Create a new socket level option to report number of queue overflows
Recently I augmented the AF_PACKET protocol to report the number of frames lost
on the socket receive queue between any two enqueued frames. This value was
exported via a SOL_PACKET level cmsg. AFter I completed that work it was
requested that this feature be generalized so that any datagram oriented socket
could make use of this option. As such I've created this patch, It creates a
new SOL_SOCKET level option called SO_RXQ_OVFL, which when enabled exports a
SOL_SOCKET level cmsg that reports the nubmer of times the sk_receive_queue
overflowed between any two given frames. It also augments the AF_PACKET
protocol to take advantage of this new feature (as it previously did not touch
sk->sk_drops, which this patch uses to record the overflow count). Tested
successfully by me.
Notes:
1) Unlike my previous patch, this patch simply records the sk_drops value, which
is not a number of drops between packets, but rather a total number of drops.
Deltas must be computed in user space.
2) While this patch currently works with datagram oriented protocols, it will
also be accepted by non-datagram oriented protocols. I'm not sure if thats
agreeable to everyone, but my argument in favor of doing so is that, for those
protocols which aren't applicable to this option, sk_drops will always be zero,
and reporting no drops on a receive queue that isn't used for those
non-participating protocols seems reasonable to me. This also saves us having
to code in a per-protocol opt in mechanism.
3) This applies cleanly to net-next assuming that commit
977750076d98c7ff6cbda51858bb5a5894a9d9ab (my af packet cmsg patch) is reverted
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-10-12 14:26:31 -06:00
|
|
|
skb->dropcount = atomic_read(&sk->sk_drops);
|
2005-04-16 16:20:36 -06:00
|
|
|
__skb_queue_tail(&sk->sk_receive_queue, skb);
|
|
|
|
spin_unlock(&sk->sk_receive_queue.lock);
|
|
|
|
sk->sk_data_ready(sk, skb->len);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
drop_n_acct:
|
make PACKET_STATISTICS getsockopt report consistently between ring and non-ring
This is a minor change.
Up until kernel 2.6.32, getsockopt(fd, SOL_PACKET, PACKET_STATISTICS,
...) would return total and dropped packets since its last invocation. The
introduction of socket queue overflow reporting [1] changed drop
rate calculation in the normal packet socket path, but not when using a
packet ring. As a result, the getsockopt now returns different statistics
depending on the reception method used. With a ring, it still returns the
count since the last call, as counts are incremented in tpacket_rcv and
reset in getsockopt. Without a ring, it returns 0 if no drops occurred
since the last getsockopt and the total drops over the lifespan of
the socket otherwise. The culprit is this line in packet_rcv, executed
on a drop:
drop_n_acct:
po->stats.tp_drops = atomic_inc_return(&sk->sk_drops);
As it shows, the new drop number it taken from the socket drop counter,
which is not reset at getsockopt. I put together a small example
that demonstrates the issue [2]. It runs for 10 seconds and overflows
the queue/ring on every odd second. The reported drop rates are:
ring: 16, 0, 16, 0, 16, ...
non-ring: 0, 15, 0, 30, 0, 46, 0, 60, 0 , 74.
Note how the even ring counts monotonically increase. Because the
getsockopt adds tp_drops to tp_packets, total counts are similarly
reported cumulatively. Long story short, reinstating the original code, as
the below patch does, fixes the issue at the cost of additional per-packet
cycles. Another solution that does not introduce per-packet overhead
is be to keep the current data path, record the value of sk_drops at
getsockopt() at call N in a new field in struct packetsock and subtract
that when reporting at call N+1. I'll be happy to code that, instead,
it's just more messy.
[1] http://patchwork.ozlabs.org/patch/35665/
[2] http://kernel.googlecode.com/files/test-packetsock-getstatistics.c
Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2011-09-30 04:38:28 -06:00
|
|
|
spin_lock(&sk->sk_receive_queue.lock);
|
|
|
|
po->stats.tp_drops++;
|
|
|
|
atomic_inc(&sk->sk_drops);
|
|
|
|
spin_unlock(&sk->sk_receive_queue.lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
drop_n_restore:
|
|
|
|
if (skb_head != skb->data && skb_shared(skb)) {
|
|
|
|
skb->data = skb_head;
|
|
|
|
skb->len = skb_len;
|
|
|
|
}
|
|
|
|
drop:
|
2009-03-11 03:49:55 -06:00
|
|
|
consume_skb(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev,
|
|
|
|
struct packet_type *pt, struct net_device *orig_dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk;
|
|
|
|
struct packet_sock *po;
|
|
|
|
struct sockaddr_ll *sll;
|
2008-07-14 23:50:15 -06:00
|
|
|
union {
|
|
|
|
struct tpacket_hdr *h1;
|
|
|
|
struct tpacket2_hdr *h2;
|
2011-08-19 04:18:16 -06:00
|
|
|
struct tpacket3_hdr *h3;
|
2008-07-14 23:50:15 -06:00
|
|
|
void *raw;
|
|
|
|
} h;
|
2009-07-21 15:57:59 -06:00
|
|
|
u8 *skb_head = skb->data;
|
2005-04-16 16:20:36 -06:00
|
|
|
int skb_len = skb->len;
|
2007-01-24 16:21:02 -07:00
|
|
|
unsigned int snaplen, res;
|
2011-08-19 04:18:16 -06:00
|
|
|
unsigned long status = TP_STATUS_USER;
|
2008-07-14 23:50:15 -06:00
|
|
|
unsigned short macoff, netoff, hdrlen;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct sk_buff *copy_skb = NULL;
|
2007-04-19 17:16:32 -06:00
|
|
|
struct timeval tv;
|
2008-07-14 23:50:15 -06:00
|
|
|
struct timespec ts;
|
2010-06-02 06:53:56 -06:00
|
|
|
struct skb_shared_hwtstamps *shhwtstamps = skb_hwtstamps(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (skb->pkt_type == PACKET_LOOPBACK)
|
|
|
|
goto drop;
|
|
|
|
|
|
|
|
sk = pt->af_packet_priv;
|
|
|
|
po = pkt_sk(sk);
|
|
|
|
|
2009-11-25 16:14:13 -07:00
|
|
|
if (!net_eq(dev_net(dev), sock_net(sk)))
|
2007-11-19 23:28:35 -07:00
|
|
|
goto drop;
|
|
|
|
|
2007-10-09 02:40:57 -06:00
|
|
|
if (dev->header_ops) {
|
2005-04-16 16:20:36 -06:00
|
|
|
if (sk->sk_type != SOCK_DGRAM)
|
2007-03-19 16:33:04 -06:00
|
|
|
skb_push(skb, skb->data - skb_mac_header(skb));
|
2005-04-16 16:20:36 -06:00
|
|
|
else if (skb->pkt_type == PACKET_OUTGOING) {
|
|
|
|
/* Special case: outgoing packets have ll header at head */
|
2007-03-10 18:16:10 -07:00
|
|
|
skb_pull(skb, skb_network_offset(skb));
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2007-02-05 00:31:32 -07:00
|
|
|
if (skb->ip_summed == CHECKSUM_PARTIAL)
|
|
|
|
status |= TP_STATUS_CSUMNOTREADY;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
snaplen = skb->len;
|
|
|
|
|
2007-01-24 16:21:02 -07:00
|
|
|
res = run_filter(skb, sk, snaplen);
|
|
|
|
if (!res)
|
2006-08-31 16:28:39 -06:00
|
|
|
goto drop_n_restore;
|
2007-01-24 16:21:02 -07:00
|
|
|
if (snaplen > res)
|
|
|
|
snaplen = res;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (sk->sk_type == SOCK_DGRAM) {
|
2008-07-18 19:05:19 -06:00
|
|
|
macoff = netoff = TPACKET_ALIGN(po->tp_hdrlen) + 16 +
|
|
|
|
po->tp_reserve;
|
2005-04-16 16:20:36 -06:00
|
|
|
} else {
|
2007-03-10 18:16:10 -07:00
|
|
|
unsigned maclen = skb_network_offset(skb);
|
2008-07-14 23:50:15 -06:00
|
|
|
netoff = TPACKET_ALIGN(po->tp_hdrlen +
|
2008-07-18 19:05:19 -06:00
|
|
|
(maclen < 16 ? 16 : maclen)) +
|
|
|
|
po->tp_reserve;
|
2005-04-16 16:20:36 -06:00
|
|
|
macoff = netoff - maclen;
|
|
|
|
}
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version <= TPACKET_V2) {
|
|
|
|
if (macoff + snaplen > po->rx_ring.frame_size) {
|
|
|
|
if (po->copy_thresh &&
|
2011-12-21 00:11:44 -07:00
|
|
|
atomic_read(&sk->sk_rmem_alloc) < sk->sk_rcvbuf) {
|
2011-08-19 04:18:16 -06:00
|
|
|
if (skb_shared(skb)) {
|
|
|
|
copy_skb = skb_clone(skb, GFP_ATOMIC);
|
|
|
|
} else {
|
|
|
|
copy_skb = skb_get(skb);
|
|
|
|
skb_head = skb->data;
|
|
|
|
}
|
|
|
|
if (copy_skb)
|
|
|
|
skb_set_owner_r(copy_skb, sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2011-08-19 04:18:16 -06:00
|
|
|
snaplen = po->rx_ring.frame_size - macoff;
|
|
|
|
if ((int)snaplen < 0)
|
|
|
|
snaplen = 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
spin_lock(&sk->sk_receive_queue.lock);
|
2011-08-19 04:18:16 -06:00
|
|
|
h.raw = packet_current_rx_frame(po, skb,
|
|
|
|
TP_STATUS_KERNEL, (macoff+snaplen));
|
2008-07-14 23:50:15 -06:00
|
|
|
if (!h.raw)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto ring_is_full;
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version <= TPACKET_V2) {
|
|
|
|
packet_increment_rx_head(po, &po->rx_ring);
|
|
|
|
/*
|
|
|
|
* LOSING will be reported till you read the stats,
|
|
|
|
* because it's COR - Clear On Read.
|
|
|
|
* Anyways, moving it for V1/V2 only as V3 doesn't need this
|
|
|
|
* at packet level.
|
|
|
|
*/
|
|
|
|
if (po->stats.tp_drops)
|
|
|
|
status |= TP_STATUS_LOSING;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
po->stats.tp_packets++;
|
|
|
|
if (copy_skb) {
|
|
|
|
status |= TP_STATUS_COPY;
|
|
|
|
__skb_queue_tail(&sk->sk_receive_queue, copy_skb);
|
|
|
|
}
|
|
|
|
spin_unlock(&sk->sk_receive_queue.lock);
|
|
|
|
|
2008-07-14 23:50:15 -06:00
|
|
|
skb_copy_bits(skb, 0, h.raw + macoff, snaplen);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2008-07-14 23:50:15 -06:00
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
h.h1->tp_len = skb->len;
|
|
|
|
h.h1->tp_snaplen = snaplen;
|
|
|
|
h.h1->tp_mac = macoff;
|
|
|
|
h.h1->tp_net = netoff;
|
2010-06-02 06:53:56 -06:00
|
|
|
if ((po->tp_tstamp & SOF_TIMESTAMPING_SYS_HARDWARE)
|
|
|
|
&& shhwtstamps->syststamp.tv64)
|
|
|
|
tv = ktime_to_timeval(shhwtstamps->syststamp);
|
|
|
|
else if ((po->tp_tstamp & SOF_TIMESTAMPING_RAW_HARDWARE)
|
|
|
|
&& shhwtstamps->hwtstamp.tv64)
|
|
|
|
tv = ktime_to_timeval(shhwtstamps->hwtstamp);
|
|
|
|
else if (skb->tstamp.tv64)
|
2008-07-14 23:50:15 -06:00
|
|
|
tv = ktime_to_timeval(skb->tstamp);
|
|
|
|
else
|
|
|
|
do_gettimeofday(&tv);
|
|
|
|
h.h1->tp_sec = tv.tv_sec;
|
|
|
|
h.h1->tp_usec = tv.tv_usec;
|
|
|
|
hdrlen = sizeof(*h.h1);
|
|
|
|
break;
|
|
|
|
case TPACKET_V2:
|
|
|
|
h.h2->tp_len = skb->len;
|
|
|
|
h.h2->tp_snaplen = snaplen;
|
|
|
|
h.h2->tp_mac = macoff;
|
|
|
|
h.h2->tp_net = netoff;
|
2010-06-02 06:53:56 -06:00
|
|
|
if ((po->tp_tstamp & SOF_TIMESTAMPING_SYS_HARDWARE)
|
|
|
|
&& shhwtstamps->syststamp.tv64)
|
|
|
|
ts = ktime_to_timespec(shhwtstamps->syststamp);
|
|
|
|
else if ((po->tp_tstamp & SOF_TIMESTAMPING_RAW_HARDWARE)
|
|
|
|
&& shhwtstamps->hwtstamp.tv64)
|
|
|
|
ts = ktime_to_timespec(shhwtstamps->hwtstamp);
|
|
|
|
else if (skb->tstamp.tv64)
|
2008-07-14 23:50:15 -06:00
|
|
|
ts = ktime_to_timespec(skb->tstamp);
|
|
|
|
else
|
|
|
|
getnstimeofday(&ts);
|
|
|
|
h.h2->tp_sec = ts.tv_sec;
|
|
|
|
h.h2->tp_nsec = ts.tv_nsec;
|
2011-06-01 00:49:10 -06:00
|
|
|
if (vlan_tx_tag_present(skb)) {
|
|
|
|
h.h2->tp_vlan_tci = vlan_tx_tag_get(skb);
|
|
|
|
status |= TP_STATUS_VLAN_VALID;
|
|
|
|
} else {
|
|
|
|
h.h2->tp_vlan_tci = 0;
|
|
|
|
}
|
2011-06-06 23:42:06 -06:00
|
|
|
h.h2->tp_padding = 0;
|
2008-07-14 23:50:15 -06:00
|
|
|
hdrlen = sizeof(*h.h2);
|
|
|
|
break;
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
|
|
|
/* tp_nxt_offset,vlan are already populated above.
|
|
|
|
* So DONT clear those fields here
|
|
|
|
*/
|
|
|
|
h.h3->tp_status |= status;
|
|
|
|
h.h3->tp_len = skb->len;
|
|
|
|
h.h3->tp_snaplen = snaplen;
|
|
|
|
h.h3->tp_mac = macoff;
|
|
|
|
h.h3->tp_net = netoff;
|
|
|
|
if ((po->tp_tstamp & SOF_TIMESTAMPING_SYS_HARDWARE)
|
|
|
|
&& shhwtstamps->syststamp.tv64)
|
|
|
|
ts = ktime_to_timespec(shhwtstamps->syststamp);
|
|
|
|
else if ((po->tp_tstamp & SOF_TIMESTAMPING_RAW_HARDWARE)
|
|
|
|
&& shhwtstamps->hwtstamp.tv64)
|
|
|
|
ts = ktime_to_timespec(shhwtstamps->hwtstamp);
|
|
|
|
else if (skb->tstamp.tv64)
|
|
|
|
ts = ktime_to_timespec(skb->tstamp);
|
|
|
|
else
|
|
|
|
getnstimeofday(&ts);
|
|
|
|
h.h3->tp_sec = ts.tv_sec;
|
|
|
|
h.h3->tp_nsec = ts.tv_nsec;
|
|
|
|
hdrlen = sizeof(*h.h3);
|
|
|
|
break;
|
2008-07-14 23:50:15 -06:00
|
|
|
default:
|
|
|
|
BUG();
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2008-07-14 23:50:15 -06:00
|
|
|
sll = h.raw + TPACKET_ALIGN(hdrlen);
|
2007-09-26 23:13:38 -06:00
|
|
|
sll->sll_halen = dev_parse_header(skb, sll->sll_addr);
|
2005-04-16 16:20:36 -06:00
|
|
|
sll->sll_family = AF_PACKET;
|
|
|
|
sll->sll_hatype = dev->type;
|
|
|
|
sll->sll_protocol = skb->protocol;
|
|
|
|
sll->sll_pkttype = skb->pkt_type;
|
2007-11-10 23:03:25 -07:00
|
|
|
if (unlikely(po->origdev))
|
2007-04-20 17:05:39 -06:00
|
|
|
sll->sll_ifindex = orig_dev->ifindex;
|
|
|
|
else
|
|
|
|
sll->sll_ifindex = dev->ifindex;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-12-07 01:11:33 -07:00
|
|
|
smp_mb();
|
2010-12-06 21:26:16 -07:00
|
|
|
#if ARCH_IMPLEMENTS_FLUSH_DCACHE_PAGE == 1
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2010-11-30 19:52:20 -07:00
|
|
|
u8 *start, *end;
|
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version <= TPACKET_V2) {
|
|
|
|
end = (u8 *)PAGE_ALIGN((unsigned long)h.raw
|
|
|
|
+ macoff + snaplen);
|
|
|
|
for (start = h.raw; start < end; start += PAGE_SIZE)
|
|
|
|
flush_dcache_page(pgv_to_page(start));
|
|
|
|
}
|
2011-07-14 09:36:33 -06:00
|
|
|
smp_wmb();
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2010-12-06 21:26:16 -07:00
|
|
|
#endif
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version <= TPACKET_V2)
|
|
|
|
__packet_set_status(po, h.raw, status);
|
|
|
|
else
|
|
|
|
prb_clear_blk_fill_status(&po->rx_ring);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
sk->sk_data_ready(sk, 0);
|
|
|
|
|
|
|
|
drop_n_restore:
|
|
|
|
if (skb_head != skb->data && skb_shared(skb)) {
|
|
|
|
skb->data = skb_head;
|
|
|
|
skb->len = skb_len;
|
|
|
|
}
|
|
|
|
drop:
|
2007-02-09 07:25:10 -07:00
|
|
|
kfree_skb(skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
ring_is_full:
|
|
|
|
po->stats.tp_drops++;
|
|
|
|
spin_unlock(&sk->sk_receive_queue.lock);
|
|
|
|
|
|
|
|
sk->sk_data_ready(sk, 0);
|
2009-02-24 17:36:42 -07:00
|
|
|
kfree_skb(copy_skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
goto drop_n_restore;
|
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static void tpacket_destruct_skb(struct sk_buff *skb)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(skb->sk);
|
2009-07-21 15:57:59 -06:00
|
|
|
void *ph;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
if (likely(po->tx_ring.pg_vec)) {
|
|
|
|
ph = skb_shinfo(skb)->destructor_arg;
|
|
|
|
BUG_ON(__packet_get_status(po, ph) != TP_STATUS_SENDING);
|
|
|
|
BUG_ON(atomic_read(&po->tx_ring.pending) == 0);
|
|
|
|
atomic_dec(&po->tx_ring.pending);
|
|
|
|
__packet_set_status(po, ph, TP_STATUS_AVAILABLE);
|
|
|
|
}
|
|
|
|
|
|
|
|
sock_wfree(skb);
|
|
|
|
}
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static int tpacket_fill_skb(struct packet_sock *po, struct sk_buff *skb,
|
|
|
|
void *frame, struct net_device *dev, int size_max,
|
2011-11-17 19:20:04 -07:00
|
|
|
__be16 proto, unsigned char *addr, int hlen)
|
2009-05-18 23:11:22 -06:00
|
|
|
{
|
|
|
|
union {
|
|
|
|
struct tpacket_hdr *h1;
|
|
|
|
struct tpacket2_hdr *h2;
|
|
|
|
void *raw;
|
|
|
|
} ph;
|
|
|
|
int to_write, offset, len, tp_len, nr_frags, len_max;
|
|
|
|
struct socket *sock = po->sk.sk_socket;
|
|
|
|
struct page *page;
|
|
|
|
void *data;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
ph.raw = frame;
|
|
|
|
|
|
|
|
skb->protocol = proto;
|
|
|
|
skb->dev = dev;
|
|
|
|
skb->priority = po->sk.sk_priority;
|
2009-10-01 13:14:46 -06:00
|
|
|
skb->mark = po->sk.sk_mark;
|
2009-05-18 23:11:22 -06:00
|
|
|
skb_shinfo(skb)->destructor_arg = ph.raw;
|
|
|
|
|
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V2:
|
|
|
|
tp_len = ph.h2->tp_len;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
tp_len = ph.h1->tp_len;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (unlikely(tp_len > size_max)) {
|
2009-07-21 15:57:59 -06:00
|
|
|
pr_err("packet size is too long (%d > %d)\n", tp_len, size_max);
|
2009-05-18 23:11:22 -06:00
|
|
|
return -EMSGSIZE;
|
|
|
|
}
|
|
|
|
|
2011-11-17 19:20:04 -07:00
|
|
|
skb_reserve(skb, hlen);
|
2009-05-18 23:11:22 -06:00
|
|
|
skb_reset_network_header(skb);
|
|
|
|
|
|
|
|
data = ph.raw + po->tp_hdrlen - sizeof(struct sockaddr_ll);
|
|
|
|
to_write = tp_len;
|
|
|
|
|
|
|
|
if (sock->type == SOCK_DGRAM) {
|
|
|
|
err = dev_hard_header(skb, dev, ntohs(proto), addr,
|
|
|
|
NULL, tp_len);
|
|
|
|
if (unlikely(err < 0))
|
|
|
|
return -EINVAL;
|
2009-07-21 15:57:59 -06:00
|
|
|
} else if (dev->hard_header_len) {
|
2009-05-18 23:11:22 -06:00
|
|
|
/* net device doesn't like empty head */
|
|
|
|
if (unlikely(tp_len <= dev->hard_header_len)) {
|
2009-07-21 15:57:59 -06:00
|
|
|
pr_err("packet size is too short (%d < %d)\n",
|
|
|
|
tp_len, dev->hard_header_len);
|
2009-05-18 23:11:22 -06:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
skb_push(skb, dev->hard_header_len);
|
|
|
|
err = skb_store_bits(skb, 0, data,
|
|
|
|
dev->hard_header_len);
|
|
|
|
if (unlikely(err))
|
|
|
|
return err;
|
|
|
|
|
|
|
|
data += dev->hard_header_len;
|
|
|
|
to_write -= dev->hard_header_len;
|
|
|
|
}
|
|
|
|
|
|
|
|
err = -EFAULT;
|
|
|
|
offset = offset_in_page(data);
|
|
|
|
len_max = PAGE_SIZE - offset;
|
|
|
|
len = ((to_write > len_max) ? len_max : to_write);
|
|
|
|
|
|
|
|
skb->data_len = to_write;
|
|
|
|
skb->len += to_write;
|
|
|
|
skb->truesize += to_write;
|
|
|
|
atomic_add(to_write, &po->sk.sk_wmem_alloc);
|
|
|
|
|
|
|
|
while (likely(to_write)) {
|
|
|
|
nr_frags = skb_shinfo(skb)->nr_frags;
|
|
|
|
|
|
|
|
if (unlikely(nr_frags >= MAX_SKB_FRAGS)) {
|
2009-07-21 15:57:59 -06:00
|
|
|
pr_err("Packet exceed the number of skb frags(%lu)\n",
|
|
|
|
MAX_SKB_FRAGS);
|
2009-05-18 23:11:22 -06:00
|
|
|
return -EFAULT;
|
|
|
|
}
|
|
|
|
|
2010-11-30 19:52:20 -07:00
|
|
|
page = pgv_to_page(data);
|
|
|
|
data += len;
|
2009-05-18 23:11:22 -06:00
|
|
|
flush_dcache_page(page);
|
|
|
|
get_page(page);
|
2010-11-30 19:52:20 -07:00
|
|
|
skb_fill_page_desc(skb, nr_frags, page, offset, len);
|
2009-05-18 23:11:22 -06:00
|
|
|
to_write -= len;
|
|
|
|
offset = 0;
|
|
|
|
len_max = PAGE_SIZE;
|
|
|
|
len = ((to_write > len_max) ? len_max : to_write);
|
|
|
|
}
|
|
|
|
|
|
|
|
return tp_len;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int tpacket_snd(struct packet_sock *po, struct msghdr *msg)
|
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
|
|
|
struct net_device *dev;
|
|
|
|
__be16 proto;
|
2011-06-01 01:18:53 -06:00
|
|
|
bool need_rls_dev = false;
|
|
|
|
int err, reserve = 0;
|
2009-07-21 15:57:59 -06:00
|
|
|
void *ph;
|
|
|
|
struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
|
2009-05-18 23:11:22 -06:00
|
|
|
int tp_len, size_max;
|
|
|
|
unsigned char *addr;
|
|
|
|
int len_sum = 0;
|
|
|
|
int status = 0;
|
2011-11-17 19:20:04 -07:00
|
|
|
int hlen, tlen;
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
mutex_lock(&po->pg_vec_lock);
|
|
|
|
|
|
|
|
err = -EBUSY;
|
|
|
|
if (saddr == NULL) {
|
2011-06-01 01:18:53 -06:00
|
|
|
dev = po->prot_hook.dev;
|
2009-05-18 23:11:22 -06:00
|
|
|
proto = po->num;
|
|
|
|
addr = NULL;
|
|
|
|
} else {
|
|
|
|
err = -EINVAL;
|
|
|
|
if (msg->msg_namelen < sizeof(struct sockaddr_ll))
|
|
|
|
goto out;
|
|
|
|
if (msg->msg_namelen < (saddr->sll_halen
|
|
|
|
+ offsetof(struct sockaddr_ll,
|
|
|
|
sll_addr)))
|
|
|
|
goto out;
|
|
|
|
proto = saddr->sll_protocol;
|
|
|
|
addr = saddr->sll_addr;
|
2011-06-01 01:18:53 -06:00
|
|
|
dev = dev_get_by_index(sock_net(&po->sk), saddr->sll_ifindex);
|
|
|
|
need_rls_dev = true;
|
2009-05-18 23:11:22 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
err = -ENXIO;
|
|
|
|
if (unlikely(dev == NULL))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
reserve = dev->hard_header_len;
|
|
|
|
|
|
|
|
err = -ENETDOWN;
|
|
|
|
if (unlikely(!(dev->flags & IFF_UP)))
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
size_max = po->tx_ring.frame_size
|
2009-10-29 04:19:11 -06:00
|
|
|
- (po->tp_hdrlen - sizeof(struct sockaddr_ll));
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
if (size_max > dev->mtu + reserve)
|
|
|
|
size_max = dev->mtu + reserve;
|
|
|
|
|
|
|
|
do {
|
|
|
|
ph = packet_current_frame(po, &po->tx_ring,
|
|
|
|
TP_STATUS_SEND_REQUEST);
|
|
|
|
|
|
|
|
if (unlikely(ph == NULL)) {
|
|
|
|
schedule();
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
status = TP_STATUS_SEND_REQUEST;
|
2011-11-17 19:20:04 -07:00
|
|
|
hlen = LL_RESERVED_SPACE(dev);
|
|
|
|
tlen = dev->needed_tailroom;
|
2009-05-18 23:11:22 -06:00
|
|
|
skb = sock_alloc_send_skb(&po->sk,
|
2011-11-17 19:20:04 -07:00
|
|
|
hlen + tlen + sizeof(struct sockaddr_ll),
|
2009-05-18 23:11:22 -06:00
|
|
|
0, &err);
|
|
|
|
|
|
|
|
if (unlikely(skb == NULL))
|
|
|
|
goto out_status;
|
|
|
|
|
|
|
|
tp_len = tpacket_fill_skb(po, skb, ph, dev, size_max, proto,
|
2011-11-17 19:20:04 -07:00
|
|
|
addr, hlen);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
if (unlikely(tp_len < 0)) {
|
|
|
|
if (po->tp_loss) {
|
|
|
|
__packet_set_status(po, ph,
|
|
|
|
TP_STATUS_AVAILABLE);
|
|
|
|
packet_increment_head(&po->tx_ring);
|
|
|
|
kfree_skb(skb);
|
|
|
|
continue;
|
|
|
|
} else {
|
|
|
|
status = TP_STATUS_WRONG_FORMAT;
|
|
|
|
err = tp_len;
|
|
|
|
goto out_status;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
skb->destructor = tpacket_destruct_skb;
|
|
|
|
__packet_set_status(po, ph, TP_STATUS_SENDING);
|
|
|
|
atomic_inc(&po->tx_ring.pending);
|
|
|
|
|
|
|
|
status = TP_STATUS_SEND_REQUEST;
|
|
|
|
err = dev_queue_xmit(skb);
|
2010-01-10 15:04:19 -07:00
|
|
|
if (unlikely(err > 0)) {
|
|
|
|
err = net_xmit_errno(err);
|
|
|
|
if (err && __packet_get_status(po, ph) ==
|
|
|
|
TP_STATUS_AVAILABLE) {
|
|
|
|
/* skb was destructed already */
|
|
|
|
skb = NULL;
|
|
|
|
goto out_status;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* skb was dropped but not destructed yet;
|
|
|
|
* let's treat it like congestion or err < 0
|
|
|
|
*/
|
|
|
|
err = 0;
|
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
packet_increment_head(&po->tx_ring);
|
|
|
|
len_sum += tp_len;
|
2009-11-29 17:55:45 -07:00
|
|
|
} while (likely((ph != NULL) ||
|
|
|
|
((!(msg->msg_flags & MSG_DONTWAIT)) &&
|
|
|
|
(atomic_read(&po->tx_ring.pending))))
|
|
|
|
);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
err = len_sum;
|
|
|
|
goto out_put;
|
|
|
|
|
|
|
|
out_status:
|
|
|
|
__packet_set_status(po, ph, status);
|
|
|
|
kfree_skb(skb);
|
|
|
|
out_put:
|
2011-06-01 01:18:53 -06:00
|
|
|
if (need_rls_dev)
|
|
|
|
dev_put(dev);
|
2009-05-18 23:11:22 -06:00
|
|
|
out:
|
|
|
|
mutex_unlock(&po->pg_vec_lock);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static struct sk_buff *packet_alloc_skb(struct sock *sk, size_t prepad,
|
|
|
|
size_t reserve, size_t len,
|
|
|
|
size_t linear, int noblock,
|
|
|
|
int *err)
|
2010-02-04 21:24:10 -07:00
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
|
|
|
|
|
|
|
/* Under a page? Don't bother with paged skb. */
|
|
|
|
if (prepad + len < PAGE_SIZE || !linear)
|
|
|
|
linear = len;
|
|
|
|
|
|
|
|
skb = sock_alloc_send_pskb(sk, prepad + linear, len - linear, noblock,
|
|
|
|
err);
|
|
|
|
if (!skb)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
skb_reserve(skb, reserve);
|
|
|
|
skb_put(skb, linear);
|
|
|
|
skb->data_len = len - linear;
|
|
|
|
skb->len += len - linear;
|
|
|
|
|
|
|
|
return skb;
|
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static int packet_snd(struct socket *sock,
|
2005-04-16 16:20:36 -06:00
|
|
|
struct msghdr *msg, size_t len)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
2009-07-21 15:57:59 -06:00
|
|
|
struct sockaddr_ll *saddr = (struct sockaddr_ll *)msg->msg_name;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct sk_buff *skb;
|
|
|
|
struct net_device *dev;
|
2006-11-08 01:26:29 -07:00
|
|
|
__be16 proto;
|
2011-06-01 01:18:53 -06:00
|
|
|
bool need_rls_dev = false;
|
2005-04-16 16:20:36 -06:00
|
|
|
unsigned char *addr;
|
2011-06-01 01:18:53 -06:00
|
|
|
int err, reserve = 0;
|
2010-02-04 21:24:10 -07:00
|
|
|
struct virtio_net_hdr vnet_hdr = { 0 };
|
|
|
|
int offset = 0;
|
|
|
|
int vnet_hdr_len;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
unsigned short gso_type = 0;
|
2011-11-17 19:20:04 -07:00
|
|
|
int hlen, tlen;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/*
|
2007-02-09 07:25:10 -07:00
|
|
|
* Get and verify the address.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (saddr == NULL) {
|
2011-06-01 01:18:53 -06:00
|
|
|
dev = po->prot_hook.dev;
|
2005-04-16 16:20:36 -06:00
|
|
|
proto = po->num;
|
|
|
|
addr = NULL;
|
|
|
|
} else {
|
|
|
|
err = -EINVAL;
|
|
|
|
if (msg->msg_namelen < sizeof(struct sockaddr_ll))
|
|
|
|
goto out;
|
2005-09-21 01:11:37 -06:00
|
|
|
if (msg->msg_namelen < (saddr->sll_halen + offsetof(struct sockaddr_ll, sll_addr)))
|
|
|
|
goto out;
|
2005-04-16 16:20:36 -06:00
|
|
|
proto = saddr->sll_protocol;
|
|
|
|
addr = saddr->sll_addr;
|
2011-06-01 01:18:53 -06:00
|
|
|
dev = dev_get_by_index(sock_net(sk), saddr->sll_ifindex);
|
|
|
|
need_rls_dev = true;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
err = -ENXIO;
|
|
|
|
if (dev == NULL)
|
|
|
|
goto out_unlock;
|
|
|
|
if (sock->type == SOCK_RAW)
|
|
|
|
reserve = dev->hard_header_len;
|
|
|
|
|
2007-01-25 20:30:36 -07:00
|
|
|
err = -ENETDOWN;
|
|
|
|
if (!(dev->flags & IFF_UP))
|
|
|
|
goto out_unlock;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
if (po->has_vnet_hdr) {
|
|
|
|
vnet_hdr_len = sizeof(vnet_hdr);
|
|
|
|
|
|
|
|
err = -EINVAL;
|
|
|
|
if (len < vnet_hdr_len)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
len -= vnet_hdr_len;
|
|
|
|
|
|
|
|
err = memcpy_fromiovec((void *)&vnet_hdr, msg->msg_iov,
|
|
|
|
vnet_hdr_len);
|
|
|
|
if (err < 0)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
if ((vnet_hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) &&
|
|
|
|
(vnet_hdr.csum_start + vnet_hdr.csum_offset + 2 >
|
|
|
|
vnet_hdr.hdr_len))
|
|
|
|
vnet_hdr.hdr_len = vnet_hdr.csum_start +
|
|
|
|
vnet_hdr.csum_offset + 2;
|
|
|
|
|
|
|
|
err = -EINVAL;
|
|
|
|
if (vnet_hdr.hdr_len > len)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
if (vnet_hdr.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
|
|
|
|
switch (vnet_hdr.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
|
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV4:
|
|
|
|
gso_type = SKB_GSO_TCPV4;
|
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_TCPV6:
|
|
|
|
gso_type = SKB_GSO_TCPV6;
|
|
|
|
break;
|
|
|
|
case VIRTIO_NET_HDR_GSO_UDP:
|
|
|
|
gso_type = SKB_GSO_UDP;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
goto out_unlock;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (vnet_hdr.gso_type & VIRTIO_NET_HDR_GSO_ECN)
|
|
|
|
gso_type |= SKB_GSO_TCP_ECN;
|
|
|
|
|
|
|
|
if (vnet_hdr.gso_size == 0)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
err = -EMSGSIZE;
|
2011-02-11 02:35:18 -07:00
|
|
|
if (!gso_type && (len > dev->mtu + reserve + VLAN_HLEN))
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out_unlock;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
err = -ENOBUFS;
|
2011-11-17 19:20:04 -07:00
|
|
|
hlen = LL_RESERVED_SPACE(dev);
|
|
|
|
tlen = dev->needed_tailroom;
|
|
|
|
skb = packet_alloc_skb(sk, hlen + tlen, hlen, len, vnet_hdr.hdr_len,
|
2010-02-04 21:24:10 -07:00
|
|
|
msg->msg_flags & MSG_DONTWAIT, &err);
|
2009-07-21 15:57:59 -06:00
|
|
|
if (skb == NULL)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out_unlock;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
skb_set_network_header(skb, reserve);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-10-09 02:36:32 -06:00
|
|
|
err = -EINVAL;
|
|
|
|
if (sock->type == SOCK_DGRAM &&
|
2010-02-04 21:24:10 -07:00
|
|
|
(offset = dev_hard_header(skb, dev, ntohs(proto), addr, NULL, len)) < 0)
|
2007-10-09 02:36:32 -06:00
|
|
|
goto out_free;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/* Returns -EFAULT on error */
|
2010-02-04 21:24:10 -07:00
|
|
|
err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (err)
|
|
|
|
goto out_free;
|
2010-08-17 02:59:14 -06:00
|
|
|
err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
|
2010-04-07 16:41:28 -06:00
|
|
|
if (err < 0)
|
|
|
|
goto out_free;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-02-11 02:35:18 -07:00
|
|
|
if (!gso_type && (len > dev->mtu + reserve)) {
|
|
|
|
/* Earlier code assumed this would be a VLAN pkt,
|
|
|
|
* double-check this now that we have the actual
|
|
|
|
* packet in hand.
|
|
|
|
*/
|
|
|
|
struct ethhdr *ehdr;
|
|
|
|
skb_reset_mac_header(skb);
|
|
|
|
ehdr = eth_hdr(skb);
|
|
|
|
if (ehdr->h_proto != htons(ETH_P_8021Q)) {
|
|
|
|
err = -EMSGSIZE;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
skb->protocol = proto;
|
|
|
|
skb->dev = dev;
|
|
|
|
skb->priority = sk->sk_priority;
|
2009-10-01 13:14:46 -06:00
|
|
|
skb->mark = sk->sk_mark;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
if (po->has_vnet_hdr) {
|
|
|
|
if (vnet_hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
|
|
|
|
if (!skb_partial_csum_set(skb, vnet_hdr.csum_start,
|
|
|
|
vnet_hdr.csum_offset)) {
|
|
|
|
err = -EINVAL;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
skb_shinfo(skb)->gso_size = vnet_hdr.gso_size;
|
|
|
|
skb_shinfo(skb)->gso_type = gso_type;
|
|
|
|
|
|
|
|
/* Header must be checked, and gso_segs computed. */
|
|
|
|
skb_shinfo(skb)->gso_type |= SKB_GSO_DODGY;
|
|
|
|
skb_shinfo(skb)->gso_segs = 0;
|
|
|
|
|
|
|
|
len += vnet_hdr_len;
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Now send it
|
|
|
|
*/
|
|
|
|
|
|
|
|
err = dev_queue_xmit(skb);
|
|
|
|
if (err > 0 && (err = net_xmit_errno(err)) != 0)
|
|
|
|
goto out_unlock;
|
|
|
|
|
2011-06-01 01:18:53 -06:00
|
|
|
if (need_rls_dev)
|
|
|
|
dev_put(dev);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
return len;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
out_free:
|
|
|
|
kfree_skb(skb);
|
|
|
|
out_unlock:
|
2011-06-01 01:18:53 -06:00
|
|
|
if (dev && need_rls_dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
dev_put(dev);
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static int packet_sendmsg(struct kiocb *iocb, struct socket *sock,
|
|
|
|
struct msghdr *msg, size_t len)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
if (po->tx_ring.pg_vec)
|
|
|
|
return tpacket_snd(po, msg);
|
|
|
|
else
|
|
|
|
return packet_snd(sock, msg, len);
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Close a PACKET socket. This is fairly simple. We immediately go
|
|
|
|
* to 'closed' state and remove our protocol entry in the device list.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int packet_release(struct socket *sock)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po;
|
2007-11-19 23:28:35 -07:00
|
|
|
struct net *net;
|
2011-08-19 04:18:16 -06:00
|
|
|
union tpacket_req_u req_u;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (!sk)
|
|
|
|
return 0;
|
|
|
|
|
2008-03-25 11:26:21 -06:00
|
|
|
net = sock_net(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
po = pkt_sk(sk);
|
|
|
|
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_lock_bh(&net->packet.sklist_lock);
|
|
|
|
sk_del_node_init_rcu(sk);
|
2008-11-24 01:09:29 -07:00
|
|
|
sock_prot_inuse_add(net, sk->sk_prot, -1);
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_unlock_bh(&net->packet.sklist_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_lock(&po->bind_lock);
|
2011-07-04 02:44:29 -06:00
|
|
|
unregister_prot_hook(sk, false);
|
2011-06-01 01:18:52 -06:00
|
|
|
if (po->prot_hook.dev) {
|
|
|
|
dev_put(po->prot_hook.dev);
|
|
|
|
po->prot_hook.dev = NULL;
|
|
|
|
}
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_unlock(&po->bind_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
packet_flush_mclist(sk);
|
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
memset(&req_u, 0, sizeof(req_u));
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
if (po->rx_ring.pg_vec)
|
2011-08-19 04:18:16 -06:00
|
|
|
packet_set_ring(sk, &req_u, 1, 0);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
if (po->tx_ring.pg_vec)
|
2011-08-19 04:18:16 -06:00
|
|
|
packet_set_ring(sk, &req_u, 1, 1);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
fanout_release(sk);
|
|
|
|
|
2010-02-22 00:57:18 -07:00
|
|
|
synchronize_net();
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Now the socket is dead. No more input will appear.
|
|
|
|
*/
|
|
|
|
sock_orphan(sk);
|
|
|
|
sock->sk = NULL;
|
|
|
|
|
|
|
|
/* Purge queues */
|
|
|
|
|
|
|
|
skb_queue_purge(&sk->sk_receive_queue);
|
2007-11-10 22:38:48 -07:00
|
|
|
sk_refcnt_debug_release(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
sock_put(sk);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Attach a packet hook.
|
|
|
|
*/
|
|
|
|
|
2006-11-08 01:26:29 -07:00
|
|
|
static int packet_do_bind(struct sock *sk, struct net_device *dev, __be16 protocol)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2011-07-05 02:45:05 -06:00
|
|
|
|
2011-12-27 20:32:41 -07:00
|
|
|
if (po->fanout) {
|
|
|
|
if (dev)
|
|
|
|
dev_put(dev);
|
|
|
|
|
2011-07-05 02:45:05 -06:00
|
|
|
return -EINVAL;
|
2011-12-27 20:32:41 -07:00
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
lock_sock(sk);
|
|
|
|
|
|
|
|
spin_lock(&po->bind_lock);
|
2011-07-04 02:44:29 -06:00
|
|
|
unregister_prot_hook(sk, true);
|
2005-04-16 16:20:36 -06:00
|
|
|
po->num = protocol;
|
|
|
|
po->prot_hook.type = protocol;
|
2011-06-01 01:18:52 -06:00
|
|
|
if (po->prot_hook.dev)
|
|
|
|
dev_put(po->prot_hook.dev);
|
2005-04-16 16:20:36 -06:00
|
|
|
po->prot_hook.dev = dev;
|
|
|
|
|
|
|
|
po->ifindex = dev ? dev->ifindex : 0;
|
|
|
|
|
|
|
|
if (protocol == 0)
|
|
|
|
goto out_unlock;
|
|
|
|
|
2007-11-12 22:05:20 -07:00
|
|
|
if (!dev || (dev->flags & IFF_UP)) {
|
2011-07-04 02:44:29 -06:00
|
|
|
register_prot_hook(sk);
|
2007-11-12 22:05:20 -07:00
|
|
|
} else {
|
|
|
|
sk->sk_err = ENETDOWN;
|
|
|
|
if (!sock_flag(sk, SOCK_DEAD))
|
|
|
|
sk->sk_error_report(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
out_unlock:
|
|
|
|
spin_unlock(&po->bind_lock);
|
|
|
|
release_sock(sk);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Bind a packet socket to a device
|
|
|
|
*/
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static int packet_bind_spkt(struct socket *sock, struct sockaddr *uaddr,
|
|
|
|
int addr_len)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2009-07-21 15:57:59 -06:00
|
|
|
struct sock *sk = sock->sk;
|
2005-04-16 16:20:36 -06:00
|
|
|
char name[15];
|
|
|
|
struct net_device *dev;
|
|
|
|
int err = -ENODEV;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Check legality
|
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2006-01-23 17:28:02 -07:00
|
|
|
if (addr_len != sizeof(struct sockaddr))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EINVAL;
|
2009-07-21 15:57:59 -06:00
|
|
|
strlcpy(name, uaddr->sa_data, sizeof(name));
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2008-03-25 11:26:21 -06:00
|
|
|
dev = dev_get_by_name(sock_net(sk), name);
|
2011-06-01 01:18:52 -06:00
|
|
|
if (dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
err = packet_do_bind(sk, dev, pkt_sk(sk)->num);
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int packet_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
|
|
|
|
{
|
2009-07-21 15:57:59 -06:00
|
|
|
struct sockaddr_ll *sll = (struct sockaddr_ll *)uaddr;
|
|
|
|
struct sock *sk = sock->sk;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct net_device *dev = NULL;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check legality
|
|
|
|
*/
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (addr_len < sizeof(struct sockaddr_ll))
|
|
|
|
return -EINVAL;
|
|
|
|
if (sll->sll_family != AF_PACKET)
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
if (sll->sll_ifindex) {
|
|
|
|
err = -ENODEV;
|
2008-03-25 11:26:21 -06:00
|
|
|
dev = dev_get_by_index(sock_net(sk), sll->sll_ifindex);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (dev == NULL)
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
err = packet_do_bind(sk, dev, sll->sll_protocol ? : pkt_sk(sk)->num);
|
|
|
|
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct proto packet_proto = {
|
|
|
|
.name = "PACKET",
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.obj_size = sizeof(struct packet_sock),
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
2007-02-09 07:25:10 -07:00
|
|
|
* Create a packet of type SOCK_PACKET.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
|
2009-11-05 23:18:14 -07:00
|
|
|
static int packet_create(struct net *net, struct socket *sock, int protocol,
|
|
|
|
int kern)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk;
|
|
|
|
struct packet_sock *po;
|
2006-11-08 01:26:29 -07:00
|
|
|
__be16 proto = (__force __be16)protocol; /* weird, but documented */
|
2005-04-16 16:20:36 -06:00
|
|
|
int err;
|
|
|
|
|
|
|
|
if (!capable(CAP_NET_RAW))
|
|
|
|
return -EPERM;
|
2007-05-29 14:16:31 -06:00
|
|
|
if (sock->type != SOCK_DGRAM && sock->type != SOCK_RAW &&
|
|
|
|
sock->type != SOCK_PACKET)
|
2005-04-16 16:20:36 -06:00
|
|
|
return -ESOCKTNOSUPPORT;
|
|
|
|
|
|
|
|
sock->state = SS_UNCONNECTED;
|
|
|
|
|
|
|
|
err = -ENOBUFS;
|
2007-11-01 01:39:31 -06:00
|
|
|
sk = sk_alloc(net, PF_PACKET, GFP_KERNEL, &packet_proto);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (sk == NULL)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
sock->ops = &packet_ops;
|
|
|
|
if (sock->type == SOCK_PACKET)
|
|
|
|
sock->ops = &packet_ops_spkt;
|
2007-05-29 14:16:31 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
sock_init_data(sock, sk);
|
|
|
|
|
|
|
|
po = pkt_sk(sk);
|
|
|
|
sk->sk_family = PF_PACKET;
|
2006-11-08 01:26:29 -07:00
|
|
|
po->num = proto;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
sk->sk_destruct = packet_sock_destruct;
|
2007-11-10 22:38:48 -07:00
|
|
|
sk_refcnt_debug_inc(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Attach a protocol block
|
|
|
|
*/
|
|
|
|
|
|
|
|
spin_lock_init(&po->bind_lock);
|
2009-01-30 15:12:06 -07:00
|
|
|
mutex_init(&po->pg_vec_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
po->prot_hook.func = packet_rcv;
|
2007-05-29 14:16:31 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (sock->type == SOCK_PACKET)
|
|
|
|
po->prot_hook.func = packet_rcv_spkt;
|
2007-05-29 14:16:31 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
po->prot_hook.af_packet_priv = sk;
|
|
|
|
|
2006-11-08 01:26:29 -07:00
|
|
|
if (proto) {
|
|
|
|
po->prot_hook.type = proto;
|
2011-07-04 02:44:29 -06:00
|
|
|
register_prot_hook(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_lock_bh(&net->packet.sklist_lock);
|
|
|
|
sk_add_node_rcu(sk, &net->packet.sklist);
|
2008-11-19 15:25:35 -07:00
|
|
|
sock_prot_inuse_add(net, &packet_proto, 1);
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_unlock_bh(&net->packet.sklist_lock);
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
return 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2010-04-07 16:41:28 -06:00
|
|
|
static int packet_recv_error(struct sock *sk, struct msghdr *msg, int len)
|
|
|
|
{
|
|
|
|
struct sock_exterr_skb *serr;
|
|
|
|
struct sk_buff *skb, *skb2;
|
|
|
|
int copied, err;
|
|
|
|
|
|
|
|
err = -EAGAIN;
|
|
|
|
skb = skb_dequeue(&sk->sk_error_queue);
|
|
|
|
if (skb == NULL)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
copied = skb->len;
|
|
|
|
if (copied > len) {
|
|
|
|
msg->msg_flags |= MSG_TRUNC;
|
|
|
|
copied = len;
|
|
|
|
}
|
|
|
|
err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
|
|
|
|
if (err)
|
|
|
|
goto out_free_skb;
|
|
|
|
|
|
|
|
sock_recv_timestamp(msg, sk, skb);
|
|
|
|
|
|
|
|
serr = SKB_EXT_ERR(skb);
|
|
|
|
put_cmsg(msg, SOL_PACKET, PACKET_TX_TIMESTAMP,
|
|
|
|
sizeof(serr->ee), &serr->ee);
|
|
|
|
|
|
|
|
msg->msg_flags |= MSG_ERRQUEUE;
|
|
|
|
err = copied;
|
|
|
|
|
|
|
|
/* Reset and regenerate socket error */
|
|
|
|
spin_lock_bh(&sk->sk_error_queue.lock);
|
|
|
|
sk->sk_err = 0;
|
|
|
|
if ((skb2 = skb_peek(&sk->sk_error_queue)) != NULL) {
|
|
|
|
sk->sk_err = SKB_EXT_ERR(skb2)->ee.ee_errno;
|
|
|
|
spin_unlock_bh(&sk->sk_error_queue.lock);
|
|
|
|
sk->sk_error_report(sk);
|
|
|
|
} else
|
|
|
|
spin_unlock_bh(&sk->sk_error_queue.lock);
|
|
|
|
|
|
|
|
out_free_skb:
|
|
|
|
kfree_skb(skb);
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Pull a packet from our receive queue and hand it to the user.
|
|
|
|
* If necessary we block.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static int packet_recvmsg(struct kiocb *iocb, struct socket *sock,
|
|
|
|
struct msghdr *msg, size_t len, int flags)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct sk_buff *skb;
|
|
|
|
int copied, err;
|
2005-09-21 01:11:37 -06:00
|
|
|
struct sockaddr_ll *sll;
|
2010-02-04 21:24:10 -07:00
|
|
|
int vnet_hdr_len = 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
err = -EINVAL;
|
2010-04-07 16:41:28 -06:00
|
|
|
if (flags & ~(MSG_PEEK|MSG_DONTWAIT|MSG_TRUNC|MSG_CMSG_COMPAT|MSG_ERRQUEUE))
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
#if 0
|
|
|
|
/* What error should we return now? EUNATTACH? */
|
|
|
|
if (pkt_sk(sk)->ifindex < 0)
|
|
|
|
return -ENODEV;
|
|
|
|
#endif
|
|
|
|
|
2010-04-07 16:41:28 -06:00
|
|
|
if (flags & MSG_ERRQUEUE) {
|
|
|
|
err = packet_recv_error(sk, msg, len);
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Call the generic datagram receiver. This handles all sorts
|
|
|
|
* of horrible races and re-entrancy so we can forget about it
|
|
|
|
* in the protocol layers.
|
|
|
|
*
|
|
|
|
* Now it will return ENETDOWN, if device have just gone down,
|
|
|
|
* but then it will block.
|
|
|
|
*/
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
skb = skb_recv_datagram(sk, flags, flags & MSG_DONTWAIT, &err);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
/*
|
2007-02-09 07:25:10 -07:00
|
|
|
* An error occurred so return it. Because skb_recv_datagram()
|
2005-04-16 16:20:36 -06:00
|
|
|
* handles the blocking we don't see and worry about blocking
|
|
|
|
* retries.
|
|
|
|
*/
|
|
|
|
|
2006-01-23 17:28:02 -07:00
|
|
|
if (skb == NULL)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
if (pkt_sk(sk)->has_vnet_hdr) {
|
|
|
|
struct virtio_net_hdr vnet_hdr = { 0 };
|
|
|
|
|
|
|
|
err = -EINVAL;
|
|
|
|
vnet_hdr_len = sizeof(vnet_hdr);
|
2010-11-08 04:58:45 -07:00
|
|
|
if (len < vnet_hdr_len)
|
2010-02-04 21:24:10 -07:00
|
|
|
goto out_free;
|
|
|
|
|
2010-11-08 04:58:45 -07:00
|
|
|
len -= vnet_hdr_len;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
if (skb_is_gso(skb)) {
|
|
|
|
struct skb_shared_info *sinfo = skb_shinfo(skb);
|
|
|
|
|
|
|
|
/* This is a hint as to how much should be linear. */
|
|
|
|
vnet_hdr.hdr_len = skb_headlen(skb);
|
|
|
|
vnet_hdr.gso_size = sinfo->gso_size;
|
|
|
|
if (sinfo->gso_type & SKB_GSO_TCPV4)
|
|
|
|
vnet_hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
|
|
|
|
else if (sinfo->gso_type & SKB_GSO_TCPV6)
|
|
|
|
vnet_hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
|
|
|
|
else if (sinfo->gso_type & SKB_GSO_UDP)
|
|
|
|
vnet_hdr.gso_type = VIRTIO_NET_HDR_GSO_UDP;
|
|
|
|
else if (sinfo->gso_type & SKB_GSO_FCOE)
|
|
|
|
goto out_free;
|
|
|
|
else
|
|
|
|
BUG();
|
|
|
|
if (sinfo->gso_type & SKB_GSO_TCP_ECN)
|
|
|
|
vnet_hdr.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
|
|
|
|
} else
|
|
|
|
vnet_hdr.gso_type = VIRTIO_NET_HDR_GSO_NONE;
|
|
|
|
|
|
|
|
if (skb->ip_summed == CHECKSUM_PARTIAL) {
|
|
|
|
vnet_hdr.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
|
2010-12-14 08:24:08 -07:00
|
|
|
vnet_hdr.csum_start = skb_checksum_start_offset(skb);
|
2010-02-04 21:24:10 -07:00
|
|
|
vnet_hdr.csum_offset = skb->csum_offset;
|
2011-06-09 18:56:17 -06:00
|
|
|
} else if (skb->ip_summed == CHECKSUM_UNNECESSARY) {
|
|
|
|
vnet_hdr.flags = VIRTIO_NET_HDR_F_DATA_VALID;
|
2010-02-04 21:24:10 -07:00
|
|
|
} /* else everything is zero */
|
|
|
|
|
|
|
|
err = memcpy_toiovec(msg->msg_iov, (void *)&vnet_hdr,
|
|
|
|
vnet_hdr_len);
|
|
|
|
if (err < 0)
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
2005-09-21 01:11:37 -06:00
|
|
|
/*
|
|
|
|
* If the address length field is there to be filled in, we fill
|
|
|
|
* it in now.
|
|
|
|
*/
|
|
|
|
|
2007-02-05 00:33:10 -07:00
|
|
|
sll = &PACKET_SKB_CB(skb)->sa.ll;
|
2005-09-21 01:11:37 -06:00
|
|
|
if (sock->type == SOCK_PACKET)
|
|
|
|
msg->msg_namelen = sizeof(struct sockaddr_pkt);
|
|
|
|
else
|
|
|
|
msg->msg_namelen = sll->sll_halen + offsetof(struct sockaddr_ll, sll_addr);
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* You lose any data beyond the buffer you gave. If it worries a
|
|
|
|
* user program they can ask the device for its MTU anyway.
|
|
|
|
*/
|
|
|
|
|
|
|
|
copied = skb->len;
|
2009-07-21 15:57:59 -06:00
|
|
|
if (copied > len) {
|
|
|
|
copied = len;
|
|
|
|
msg->msg_flags |= MSG_TRUNC;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
err = skb_copy_datagram_iovec(skb, 0, msg->msg_iov, copied);
|
|
|
|
if (err)
|
|
|
|
goto out_free;
|
|
|
|
|
net: Generalize socket rx gap / receive queue overflow cmsg
Create a new socket level option to report number of queue overflows
Recently I augmented the AF_PACKET protocol to report the number of frames lost
on the socket receive queue between any two enqueued frames. This value was
exported via a SOL_PACKET level cmsg. AFter I completed that work it was
requested that this feature be generalized so that any datagram oriented socket
could make use of this option. As such I've created this patch, It creates a
new SOL_SOCKET level option called SO_RXQ_OVFL, which when enabled exports a
SOL_SOCKET level cmsg that reports the nubmer of times the sk_receive_queue
overflowed between any two given frames. It also augments the AF_PACKET
protocol to take advantage of this new feature (as it previously did not touch
sk->sk_drops, which this patch uses to record the overflow count). Tested
successfully by me.
Notes:
1) Unlike my previous patch, this patch simply records the sk_drops value, which
is not a number of drops between packets, but rather a total number of drops.
Deltas must be computed in user space.
2) While this patch currently works with datagram oriented protocols, it will
also be accepted by non-datagram oriented protocols. I'm not sure if thats
agreeable to everyone, but my argument in favor of doing so is that, for those
protocols which aren't applicable to this option, sk_drops will always be zero,
and reporting no drops on a receive queue that isn't used for those
non-participating protocols seems reasonable to me. This also saves us having
to code in a per-protocol opt in mechanism.
3) This applies cleanly to net-next assuming that commit
977750076d98c7ff6cbda51858bb5a5894a9d9ab (my af packet cmsg patch) is reverted
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2009-10-12 14:26:31 -06:00
|
|
|
sock_recv_ts_and_drops(msg, sk, skb);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (msg->msg_name)
|
2007-02-05 00:33:10 -07:00
|
|
|
memcpy(msg->msg_name, &PACKET_SKB_CB(skb)->sa,
|
|
|
|
msg->msg_namelen);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2007-02-05 00:31:32 -07:00
|
|
|
if (pkt_sk(sk)->auxdata) {
|
2007-02-05 00:33:10 -07:00
|
|
|
struct tpacket_auxdata aux;
|
|
|
|
|
|
|
|
aux.tp_status = TP_STATUS_USER;
|
|
|
|
if (skb->ip_summed == CHECKSUM_PARTIAL)
|
|
|
|
aux.tp_status |= TP_STATUS_CSUMNOTREADY;
|
|
|
|
aux.tp_len = PACKET_SKB_CB(skb)->origlen;
|
|
|
|
aux.tp_snaplen = skb->len;
|
|
|
|
aux.tp_mac = 0;
|
2007-03-10 18:16:10 -07:00
|
|
|
aux.tp_net = skb_network_offset(skb);
|
2011-06-01 00:49:10 -06:00
|
|
|
if (vlan_tx_tag_present(skb)) {
|
|
|
|
aux.tp_vlan_tci = vlan_tx_tag_get(skb);
|
|
|
|
aux.tp_status |= TP_STATUS_VLAN_VALID;
|
|
|
|
} else {
|
|
|
|
aux.tp_vlan_tci = 0;
|
|
|
|
}
|
2011-06-06 23:42:06 -06:00
|
|
|
aux.tp_padding = 0;
|
2007-02-05 00:33:10 -07:00
|
|
|
put_cmsg(msg, SOL_PACKET, PACKET_AUXDATA, sizeof(aux), &aux);
|
2007-02-05 00:31:32 -07:00
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* Free or return the buffer as appropriate. Again this
|
|
|
|
* hides all the races and re-entrancy issues from us.
|
|
|
|
*/
|
2010-02-04 21:24:10 -07:00
|
|
|
err = vnet_hdr_len + ((flags&MSG_TRUNC) ? skb->len : copied);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
out_free:
|
|
|
|
skb_free_datagram(sk, skb);
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int packet_getname_spkt(struct socket *sock, struct sockaddr *uaddr,
|
|
|
|
int *uaddr_len, int peer)
|
|
|
|
{
|
|
|
|
struct net_device *dev;
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
|
|
|
|
if (peer)
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
|
|
|
uaddr->sa_family = AF_PACKET;
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
dev = dev_get_by_index_rcu(sock_net(sk), pkt_sk(sk)->ifindex);
|
|
|
|
if (dev)
|
2010-11-10 13:09:10 -07:00
|
|
|
strncpy(uaddr->sa_data, dev->name, 14);
|
2009-11-02 02:43:32 -07:00
|
|
|
else
|
2005-04-16 16:20:36 -06:00
|
|
|
memset(uaddr->sa_data, 0, 14);
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_unlock();
|
2005-04-16 16:20:36 -06:00
|
|
|
*uaddr_len = sizeof(*uaddr);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int packet_getname(struct socket *sock, struct sockaddr *uaddr,
|
|
|
|
int *uaddr_len, int peer)
|
|
|
|
{
|
|
|
|
struct net_device *dev;
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2009-11-07 22:51:19 -07:00
|
|
|
DECLARE_SOCKADDR(struct sockaddr_ll *, sll, uaddr);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (peer)
|
|
|
|
return -EOPNOTSUPP;
|
|
|
|
|
|
|
|
sll->sll_family = AF_PACKET;
|
|
|
|
sll->sll_ifindex = po->ifindex;
|
|
|
|
sll->sll_protocol = po->num;
|
2010-11-10 13:09:10 -07:00
|
|
|
sll->sll_pkttype = 0;
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
dev = dev_get_by_index_rcu(sock_net(sk), po->ifindex);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (dev) {
|
|
|
|
sll->sll_hatype = dev->type;
|
|
|
|
sll->sll_halen = dev->addr_len;
|
|
|
|
memcpy(sll->sll_addr, dev->dev_addr, dev->addr_len);
|
|
|
|
} else {
|
|
|
|
sll->sll_hatype = 0; /* Bad: we have no ARPHRD_UNSPEC */
|
|
|
|
sll->sll_halen = 0;
|
|
|
|
}
|
2009-11-02 02:43:32 -07:00
|
|
|
rcu_read_unlock();
|
2005-09-21 01:11:37 -06:00
|
|
|
*uaddr_len = offsetof(struct sockaddr_ll, sll_addr) + sll->sll_halen;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-07-14 21:49:46 -06:00
|
|
|
static int packet_dev_mc(struct net_device *dev, struct packet_mclist *i,
|
|
|
|
int what)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
switch (i->type) {
|
|
|
|
case PACKET_MR_MULTICAST:
|
2010-03-02 13:40:01 -07:00
|
|
|
if (i->alen != dev->addr_len)
|
|
|
|
return -EINVAL;
|
2005-04-16 16:20:36 -06:00
|
|
|
if (what > 0)
|
2010-04-01 15:22:57 -06:00
|
|
|
return dev_mc_add(dev, i->addr);
|
2005-04-16 16:20:36 -06:00
|
|
|
else
|
2010-04-01 15:22:57 -06:00
|
|
|
return dev_mc_del(dev, i->addr);
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
|
|
|
case PACKET_MR_PROMISC:
|
2008-07-14 21:49:46 -06:00
|
|
|
return dev_set_promiscuity(dev, what);
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
|
|
|
case PACKET_MR_ALLMULTI:
|
2008-07-14 21:49:46 -06:00
|
|
|
return dev_set_allmulti(dev, what);
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
2009-05-19 12:27:17 -06:00
|
|
|
case PACKET_MR_UNICAST:
|
2010-03-02 13:40:01 -07:00
|
|
|
if (i->alen != dev->addr_len)
|
|
|
|
return -EINVAL;
|
2009-05-19 12:27:17 -06:00
|
|
|
if (what > 0)
|
2010-04-01 15:22:09 -06:00
|
|
|
return dev_uc_add(dev, i->addr);
|
2009-05-19 12:27:17 -06:00
|
|
|
else
|
2010-04-01 15:22:09 -06:00
|
|
|
return dev_uc_del(dev, i->addr);
|
2009-05-19 12:27:17 -06:00
|
|
|
break;
|
2009-07-21 15:57:59 -06:00
|
|
|
default:
|
|
|
|
break;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2008-07-14 21:49:46 -06:00
|
|
|
return 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static void packet_dev_mclist(struct net_device *dev, struct packet_mclist *i, int what)
|
|
|
|
{
|
2009-07-21 15:57:59 -06:00
|
|
|
for ( ; i; i = i->next) {
|
2005-04-16 16:20:36 -06:00
|
|
|
if (i->ifindex == dev->ifindex)
|
|
|
|
packet_dev_mc(dev, i, what);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-09-21 01:11:37 -06:00
|
|
|
static int packet_mc_add(struct sock *sk, struct packet_mreq_max *mreq)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
struct packet_mclist *ml, *i;
|
|
|
|
struct net_device *dev;
|
|
|
|
int err;
|
|
|
|
|
|
|
|
rtnl_lock();
|
|
|
|
|
|
|
|
err = -ENODEV;
|
2008-03-25 11:26:21 -06:00
|
|
|
dev = __dev_get_by_index(sock_net(sk), mreq->mr_ifindex);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!dev)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
err = -EINVAL;
|
2010-03-02 13:40:01 -07:00
|
|
|
if (mreq->mr_alen > dev->addr_len)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto done;
|
|
|
|
|
|
|
|
err = -ENOBUFS;
|
2006-01-11 16:56:43 -07:00
|
|
|
i = kmalloc(sizeof(*i), GFP_KERNEL);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (i == NULL)
|
|
|
|
goto done;
|
|
|
|
|
|
|
|
err = 0;
|
|
|
|
for (ml = po->mclist; ml; ml = ml->next) {
|
|
|
|
if (ml->ifindex == mreq->mr_ifindex &&
|
|
|
|
ml->type == mreq->mr_type &&
|
|
|
|
ml->alen == mreq->mr_alen &&
|
|
|
|
memcmp(ml->addr, mreq->mr_address, ml->alen) == 0) {
|
|
|
|
ml->count++;
|
|
|
|
/* Free the new element ... */
|
|
|
|
kfree(i);
|
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
i->type = mreq->mr_type;
|
|
|
|
i->ifindex = mreq->mr_ifindex;
|
|
|
|
i->alen = mreq->mr_alen;
|
|
|
|
memcpy(i->addr, mreq->mr_address, i->alen);
|
|
|
|
i->count = 1;
|
|
|
|
i->next = po->mclist;
|
|
|
|
po->mclist = i;
|
2008-07-14 21:49:46 -06:00
|
|
|
err = packet_dev_mc(dev, i, 1);
|
|
|
|
if (err) {
|
|
|
|
po->mclist = i->next;
|
|
|
|
kfree(i);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
done:
|
|
|
|
rtnl_unlock();
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-09-21 01:11:37 -06:00
|
|
|
static int packet_mc_drop(struct sock *sk, struct packet_mreq_max *mreq)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct packet_mclist *ml, **mlp;
|
|
|
|
|
|
|
|
rtnl_lock();
|
|
|
|
|
|
|
|
for (mlp = &pkt_sk(sk)->mclist; (ml = *mlp) != NULL; mlp = &ml->next) {
|
|
|
|
if (ml->ifindex == mreq->mr_ifindex &&
|
|
|
|
ml->type == mreq->mr_type &&
|
|
|
|
ml->alen == mreq->mr_alen &&
|
|
|
|
memcmp(ml->addr, mreq->mr_address, ml->alen) == 0) {
|
|
|
|
if (--ml->count == 0) {
|
|
|
|
struct net_device *dev;
|
|
|
|
*mlp = ml->next;
|
2009-10-16 00:38:46 -06:00
|
|
|
dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
|
|
|
|
if (dev)
|
2005-04-16 16:20:36 -06:00
|
|
|
packet_dev_mc(dev, ml, -1);
|
|
|
|
kfree(ml);
|
|
|
|
}
|
|
|
|
rtnl_unlock();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
rtnl_unlock();
|
|
|
|
return -EADDRNOTAVAIL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void packet_flush_mclist(struct sock *sk)
|
|
|
|
{
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
struct packet_mclist *ml;
|
|
|
|
|
|
|
|
if (!po->mclist)
|
|
|
|
return;
|
|
|
|
|
|
|
|
rtnl_lock();
|
|
|
|
while ((ml = po->mclist) != NULL) {
|
|
|
|
struct net_device *dev;
|
|
|
|
|
|
|
|
po->mclist = ml->next;
|
2009-10-16 00:38:46 -06:00
|
|
|
dev = __dev_get_by_index(sock_net(sk), ml->ifindex);
|
|
|
|
if (dev != NULL)
|
2005-04-16 16:20:36 -06:00
|
|
|
packet_dev_mc(dev, ml, -1);
|
|
|
|
kfree(ml);
|
|
|
|
}
|
|
|
|
rtnl_unlock();
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2009-09-30 17:12:20 -06:00
|
|
|
packet_setsockopt(struct socket *sock, int level, int optname, char __user *optval, unsigned int optlen)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
2007-02-05 00:31:32 -07:00
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (level != SOL_PACKET)
|
|
|
|
return -ENOPROTOOPT;
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
switch (optname) {
|
2007-02-09 07:25:10 -07:00
|
|
|
case PACKET_ADD_MEMBERSHIP:
|
2005-04-16 16:20:36 -06:00
|
|
|
case PACKET_DROP_MEMBERSHIP:
|
|
|
|
{
|
2005-09-21 01:11:37 -06:00
|
|
|
struct packet_mreq_max mreq;
|
|
|
|
int len = optlen;
|
|
|
|
memset(&mreq, 0, sizeof(mreq));
|
|
|
|
if (len < sizeof(struct packet_mreq))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EINVAL;
|
2005-09-21 01:11:37 -06:00
|
|
|
if (len > sizeof(mreq))
|
|
|
|
len = sizeof(mreq);
|
2009-07-21 15:57:59 -06:00
|
|
|
if (copy_from_user(&mreq, optval, len))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EFAULT;
|
2005-09-21 01:11:37 -06:00
|
|
|
if (len < (mreq.mr_alen + offsetof(struct packet_mreq, mr_address)))
|
|
|
|
return -EINVAL;
|
2005-04-16 16:20:36 -06:00
|
|
|
if (optname == PACKET_ADD_MEMBERSHIP)
|
|
|
|
ret = packet_mc_add(sk, &mreq);
|
|
|
|
else
|
|
|
|
ret = packet_mc_drop(sk, &mreq);
|
|
|
|
return ret;
|
|
|
|
}
|
2007-05-29 14:12:50 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
case PACKET_RX_RING:
|
2009-05-18 23:11:22 -06:00
|
|
|
case PACKET_TX_RING:
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2011-08-19 04:18:16 -06:00
|
|
|
union tpacket_req_u req_u;
|
|
|
|
int len;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
case TPACKET_V2:
|
|
|
|
len = sizeof(req_u.req);
|
|
|
|
break;
|
|
|
|
case TPACKET_V3:
|
|
|
|
default:
|
|
|
|
len = sizeof(req_u.req3);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (optlen < len)
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EINVAL;
|
2010-02-04 21:24:10 -07:00
|
|
|
if (pkt_sk(sk)->has_vnet_hdr)
|
|
|
|
return -EINVAL;
|
2011-08-19 04:18:16 -06:00
|
|
|
if (copy_from_user(&req_u.req, optval, len))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EFAULT;
|
2011-08-19 04:18:16 -06:00
|
|
|
return packet_set_ring(sk, &req_u, 0,
|
|
|
|
optname == PACKET_TX_RING);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
case PACKET_COPY_THRESH:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
if (optlen != sizeof(val))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EINVAL;
|
2009-07-21 15:57:59 -06:00
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
2005-04-16 16:20:36 -06:00
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
pkt_sk(sk)->copy_thresh = val;
|
|
|
|
return 0;
|
|
|
|
}
|
2008-07-14 23:50:15 -06:00
|
|
|
case PACKET_VERSION:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (optlen != sizeof(val))
|
|
|
|
return -EINVAL;
|
2009-05-18 23:11:22 -06:00
|
|
|
if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
|
2008-07-14 23:50:15 -06:00
|
|
|
return -EBUSY;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
switch (val) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
case TPACKET_V2:
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
2008-07-14 23:50:15 -06:00
|
|
|
po->tp_version = val;
|
|
|
|
return 0;
|
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
2008-07-18 19:05:19 -06:00
|
|
|
case PACKET_RESERVE:
|
|
|
|
{
|
|
|
|
unsigned int val;
|
|
|
|
|
|
|
|
if (optlen != sizeof(val))
|
|
|
|
return -EINVAL;
|
2009-05-18 23:11:22 -06:00
|
|
|
if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
|
2008-07-18 19:05:19 -06:00
|
|
|
return -EBUSY;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
po->tp_reserve = val;
|
|
|
|
return 0;
|
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
case PACKET_LOSS:
|
|
|
|
{
|
|
|
|
unsigned int val;
|
|
|
|
|
|
|
|
if (optlen != sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
|
|
|
|
return -EBUSY;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
po->tp_loss = !!val;
|
|
|
|
return 0;
|
|
|
|
}
|
2007-02-05 00:31:32 -07:00
|
|
|
case PACKET_AUXDATA:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (optlen < sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
po->auxdata = !!val;
|
|
|
|
return 0;
|
|
|
|
}
|
2007-04-20 17:05:39 -06:00
|
|
|
case PACKET_ORIGDEV:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (optlen < sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
po->origdev = !!val;
|
|
|
|
return 0;
|
|
|
|
}
|
2010-02-04 21:24:10 -07:00
|
|
|
case PACKET_VNET_HDR:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (sock->type != SOCK_RAW)
|
|
|
|
return -EINVAL;
|
|
|
|
if (po->rx_ring.pg_vec || po->tx_ring.pg_vec)
|
|
|
|
return -EBUSY;
|
|
|
|
if (optlen < sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
po->has_vnet_hdr = !!val;
|
|
|
|
return 0;
|
|
|
|
}
|
2010-06-02 06:53:56 -06:00
|
|
|
case PACKET_TIMESTAMP:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (optlen != sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
po->tp_tstamp = val;
|
|
|
|
return 0;
|
|
|
|
}
|
2011-07-05 02:45:05 -06:00
|
|
|
case PACKET_FANOUT:
|
|
|
|
{
|
|
|
|
int val;
|
|
|
|
|
|
|
|
if (optlen != sizeof(val))
|
|
|
|
return -EINVAL;
|
|
|
|
if (copy_from_user(&val, optval, sizeof(val)))
|
|
|
|
return -EFAULT;
|
|
|
|
|
|
|
|
return fanout_add(sk, val & 0xffff, val >> 16);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
default:
|
|
|
|
return -ENOPROTOOPT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int packet_getsockopt(struct socket *sock, int level, int optname,
|
|
|
|
char __user *optval, int __user *optlen)
|
|
|
|
{
|
|
|
|
int len;
|
2007-02-05 00:31:32 -07:00
|
|
|
int val;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2007-02-05 00:31:32 -07:00
|
|
|
void *data;
|
|
|
|
struct tpacket_stats st;
|
2011-08-19 04:18:16 -06:00
|
|
|
union tpacket_stats_u st_u;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (level != SOL_PACKET)
|
|
|
|
return -ENOPROTOOPT;
|
|
|
|
|
2006-01-23 17:28:02 -07:00
|
|
|
if (get_user(len, optlen))
|
|
|
|
return -EFAULT;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
if (len < 0)
|
|
|
|
return -EINVAL;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
switch (optname) {
|
2005-04-16 16:20:36 -06:00
|
|
|
case PACKET_STATISTICS:
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version == TPACKET_V3) {
|
|
|
|
len = sizeof(struct tpacket_stats_v3);
|
|
|
|
} else {
|
|
|
|
if (len > sizeof(struct tpacket_stats))
|
|
|
|
len = sizeof(struct tpacket_stats);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
spin_lock_bh(&sk->sk_receive_queue.lock);
|
2011-08-19 04:18:16 -06:00
|
|
|
if (po->tp_version == TPACKET_V3) {
|
|
|
|
memcpy(&st_u.stats3, &po->stats,
|
|
|
|
sizeof(struct tpacket_stats));
|
|
|
|
st_u.stats3.tp_freeze_q_cnt =
|
|
|
|
po->stats_u.stats3.tp_freeze_q_cnt;
|
|
|
|
st_u.stats3.tp_packets += po->stats.tp_drops;
|
|
|
|
data = &st_u.stats3;
|
|
|
|
} else {
|
|
|
|
st = po->stats;
|
|
|
|
st.tp_packets += st.tp_drops;
|
|
|
|
data = &st;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
memset(&po->stats, 0, sizeof(st));
|
|
|
|
spin_unlock_bh(&sk->sk_receive_queue.lock);
|
2007-02-05 00:31:32 -07:00
|
|
|
break;
|
|
|
|
case PACKET_AUXDATA:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = po->auxdata;
|
|
|
|
|
2007-04-20 17:05:39 -06:00
|
|
|
data = &val;
|
|
|
|
break;
|
|
|
|
case PACKET_ORIGDEV:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = po->origdev;
|
|
|
|
|
2010-02-04 21:24:10 -07:00
|
|
|
data = &val;
|
|
|
|
break;
|
|
|
|
case PACKET_VNET_HDR:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = po->has_vnet_hdr;
|
|
|
|
|
2007-02-05 00:31:32 -07:00
|
|
|
data = &val;
|
2005-04-16 16:20:36 -06:00
|
|
|
break;
|
2008-07-14 23:50:15 -06:00
|
|
|
case PACKET_VERSION:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = po->tp_version;
|
|
|
|
data = &val;
|
|
|
|
break;
|
|
|
|
case PACKET_HDRLEN:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
if (copy_from_user(&val, optval, len))
|
|
|
|
return -EFAULT;
|
|
|
|
switch (val) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
val = sizeof(struct tpacket_hdr);
|
|
|
|
break;
|
|
|
|
case TPACKET_V2:
|
|
|
|
val = sizeof(struct tpacket2_hdr);
|
|
|
|
break;
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
|
|
|
val = sizeof(struct tpacket3_hdr);
|
|
|
|
break;
|
2008-07-14 23:50:15 -06:00
|
|
|
default:
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
data = &val;
|
|
|
|
break;
|
2008-07-18 19:05:19 -06:00
|
|
|
case PACKET_RESERVE:
|
|
|
|
if (len > sizeof(unsigned int))
|
|
|
|
len = sizeof(unsigned int);
|
|
|
|
val = po->tp_reserve;
|
|
|
|
data = &val;
|
|
|
|
break;
|
2009-05-18 23:11:22 -06:00
|
|
|
case PACKET_LOSS:
|
|
|
|
if (len > sizeof(unsigned int))
|
|
|
|
len = sizeof(unsigned int);
|
|
|
|
val = po->tp_loss;
|
|
|
|
data = &val;
|
|
|
|
break;
|
2010-06-02 06:53:56 -06:00
|
|
|
case PACKET_TIMESTAMP:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = po->tp_tstamp;
|
|
|
|
data = &val;
|
|
|
|
break;
|
2011-07-05 02:45:05 -06:00
|
|
|
case PACKET_FANOUT:
|
|
|
|
if (len > sizeof(int))
|
|
|
|
len = sizeof(int);
|
|
|
|
val = (po->fanout ?
|
|
|
|
((u32)po->fanout->id |
|
|
|
|
((u32)po->fanout->type << 16)) :
|
|
|
|
0);
|
|
|
|
data = &val;
|
|
|
|
break;
|
2005-04-16 16:20:36 -06:00
|
|
|
default:
|
|
|
|
return -ENOPROTOOPT;
|
|
|
|
}
|
|
|
|
|
2006-01-23 17:28:02 -07:00
|
|
|
if (put_user(len, optlen))
|
|
|
|
return -EFAULT;
|
2007-02-05 00:31:32 -07:00
|
|
|
if (copy_to_user(optval, data, len))
|
|
|
|
return -EFAULT;
|
2006-01-23 17:28:02 -07:00
|
|
|
return 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int packet_notifier(struct notifier_block *this, unsigned long msg, void *data)
|
|
|
|
{
|
|
|
|
struct sock *sk;
|
|
|
|
struct hlist_node *node;
|
2007-02-21 00:19:54 -07:00
|
|
|
struct net_device *dev = data;
|
2008-03-25 06:47:49 -06:00
|
|
|
struct net *net = dev_net(dev);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2010-02-22 00:57:18 -07:00
|
|
|
rcu_read_lock();
|
|
|
|
sk_for_each_rcu(sk, node, &net->packet.sklist) {
|
2005-04-16 16:20:36 -06:00
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
|
|
|
|
switch (msg) {
|
|
|
|
case NETDEV_UNREGISTER:
|
|
|
|
if (po->mclist)
|
|
|
|
packet_dev_mclist(dev, po->mclist, -1);
|
2007-05-29 14:12:50 -06:00
|
|
|
/* fallthrough */
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
case NETDEV_DOWN:
|
|
|
|
if (dev->ifindex == po->ifindex) {
|
|
|
|
spin_lock(&po->bind_lock);
|
|
|
|
if (po->running) {
|
2011-07-04 02:44:29 -06:00
|
|
|
__unregister_prot_hook(sk, false);
|
2005-04-16 16:20:36 -06:00
|
|
|
sk->sk_err = ENETDOWN;
|
|
|
|
if (!sock_flag(sk, SOCK_DEAD))
|
|
|
|
sk->sk_error_report(sk);
|
|
|
|
}
|
|
|
|
if (msg == NETDEV_UNREGISTER) {
|
|
|
|
po->ifindex = -1;
|
2011-06-01 01:18:52 -06:00
|
|
|
if (po->prot_hook.dev)
|
|
|
|
dev_put(po->prot_hook.dev);
|
2005-04-16 16:20:36 -06:00
|
|
|
po->prot_hook.dev = NULL;
|
|
|
|
}
|
|
|
|
spin_unlock(&po->bind_lock);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
case NETDEV_UP:
|
2010-02-22 00:57:18 -07:00
|
|
|
if (dev->ifindex == po->ifindex) {
|
|
|
|
spin_lock(&po->bind_lock);
|
2011-07-04 02:44:29 -06:00
|
|
|
if (po->num)
|
|
|
|
register_prot_hook(sk);
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_unlock(&po->bind_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2010-02-22 00:57:18 -07:00
|
|
|
rcu_read_unlock();
|
2005-04-16 16:20:36 -06:00
|
|
|
return NOTIFY_DONE;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int packet_ioctl(struct socket *sock, unsigned int cmd,
|
|
|
|
unsigned long arg)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
switch (cmd) {
|
2009-07-21 15:57:59 -06:00
|
|
|
case SIOCOUTQ:
|
|
|
|
{
|
|
|
|
int amount = sk_wmem_alloc_get(sk);
|
2009-06-17 20:05:41 -06:00
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
return put_user(amount, (int __user *)arg);
|
|
|
|
}
|
|
|
|
case SIOCINQ:
|
|
|
|
{
|
|
|
|
struct sk_buff *skb;
|
|
|
|
int amount = 0;
|
|
|
|
|
|
|
|
spin_lock_bh(&sk->sk_receive_queue.lock);
|
|
|
|
skb = skb_peek(&sk->sk_receive_queue);
|
|
|
|
if (skb)
|
|
|
|
amount = skb->len;
|
|
|
|
spin_unlock_bh(&sk->sk_receive_queue.lock);
|
|
|
|
return put_user(amount, (int __user *)arg);
|
|
|
|
}
|
|
|
|
case SIOCGSTAMP:
|
|
|
|
return sock_get_timestamp(sk, (struct timeval __user *)arg);
|
|
|
|
case SIOCGSTAMPNS:
|
|
|
|
return sock_get_timestampns(sk, (struct timespec __user *)arg);
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
#ifdef CONFIG_INET
|
2009-07-21 15:57:59 -06:00
|
|
|
case SIOCADDRT:
|
|
|
|
case SIOCDELRT:
|
|
|
|
case SIOCDARP:
|
|
|
|
case SIOCGARP:
|
|
|
|
case SIOCSARP:
|
|
|
|
case SIOCGIFADDR:
|
|
|
|
case SIOCSIFADDR:
|
|
|
|
case SIOCGIFBRDADDR:
|
|
|
|
case SIOCSIFBRDADDR:
|
|
|
|
case SIOCGIFNETMASK:
|
|
|
|
case SIOCSIFNETMASK:
|
|
|
|
case SIOCGIFDSTADDR:
|
|
|
|
case SIOCSIFDSTADDR:
|
|
|
|
case SIOCSIFFLAGS:
|
|
|
|
return inet_dgram_ops.ioctl(sock, cmd, arg);
|
2005-04-16 16:20:36 -06:00
|
|
|
#endif
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
default:
|
|
|
|
return -ENOIOCTLCMD;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-07-21 15:57:59 -06:00
|
|
|
static unsigned int packet_poll(struct file *file, struct socket *sock,
|
2005-04-16 16:20:36 -06:00
|
|
|
poll_table *wait)
|
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
|
|
|
unsigned int mask = datagram_poll(file, sock, wait);
|
|
|
|
|
|
|
|
spin_lock_bh(&sk->sk_receive_queue.lock);
|
2009-05-18 23:11:22 -06:00
|
|
|
if (po->rx_ring.pg_vec) {
|
2011-08-19 04:18:16 -06:00
|
|
|
if (!packet_previous_rx_frame(po, &po->rx_ring,
|
|
|
|
TP_STATUS_KERNEL))
|
2005-04-16 16:20:36 -06:00
|
|
|
mask |= POLLIN | POLLRDNORM;
|
|
|
|
}
|
|
|
|
spin_unlock_bh(&sk->sk_receive_queue.lock);
|
2009-05-18 23:11:22 -06:00
|
|
|
spin_lock_bh(&sk->sk_write_queue.lock);
|
|
|
|
if (po->tx_ring.pg_vec) {
|
|
|
|
if (packet_current_frame(po, &po->tx_ring, TP_STATUS_AVAILABLE))
|
|
|
|
mask |= POLLOUT | POLLWRNORM;
|
|
|
|
}
|
|
|
|
spin_unlock_bh(&sk->sk_write_queue.lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
return mask;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Dirty? Well, I still did not learn better way to account
|
|
|
|
* for user mmaps.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static void packet_mm_open(struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
2009-07-21 15:57:59 -06:00
|
|
|
struct socket *sock = file->private_data;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct sock *sk = sock->sk;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (sk)
|
|
|
|
atomic_inc(&pkt_sk(sk)->mapped);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void packet_mm_close(struct vm_area_struct *vma)
|
|
|
|
{
|
|
|
|
struct file *file = vma->vm_file;
|
2009-07-21 15:57:59 -06:00
|
|
|
struct socket *sock = file->private_data;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct sock *sk = sock->sk;
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
if (sk)
|
|
|
|
atomic_dec(&pkt_sk(sk)->mapped);
|
|
|
|
}
|
|
|
|
|
2009-09-27 12:29:37 -06:00
|
|
|
static const struct vm_operations_struct packet_mmap_ops = {
|
2009-07-21 15:57:59 -06:00
|
|
|
.open = packet_mm_open,
|
|
|
|
.close = packet_mm_close,
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
|
|
|
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
static void free_pg_vec(struct pgv *pg_vec, unsigned int order,
|
|
|
|
unsigned int len)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2005-12-06 17:38:35 -07:00
|
|
|
for (i = 0; i < len; i++) {
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
if (likely(pg_vec[i].buffer)) {
|
2010-11-30 19:52:57 -07:00
|
|
|
if (is_vmalloc_addr(pg_vec[i].buffer))
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
vfree(pg_vec[i].buffer);
|
|
|
|
else
|
|
|
|
free_pages((unsigned long)pg_vec[i].buffer,
|
|
|
|
order);
|
|
|
|
pg_vec[i].buffer = NULL;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
kfree(pg_vec);
|
|
|
|
}
|
|
|
|
|
2011-11-02 05:00:49 -06:00
|
|
|
static char *alloc_one_pg_vec_page(unsigned long order)
|
2005-12-06 17:38:35 -07:00
|
|
|
{
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
char *buffer = NULL;
|
|
|
|
gfp_t gfp_flags = GFP_KERNEL | __GFP_COMP |
|
|
|
|
__GFP_ZERO | __GFP_NOWARN | __GFP_NORETRY;
|
|
|
|
|
|
|
|
buffer = (char *) __get_free_pages(gfp_flags, order);
|
|
|
|
|
|
|
|
if (buffer)
|
|
|
|
return buffer;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* __get_free_pages failed, fall back to vmalloc
|
|
|
|
*/
|
2010-11-20 00:31:54 -07:00
|
|
|
buffer = vzalloc((1 << order) * PAGE_SIZE);
|
2009-04-15 04:39:52 -06:00
|
|
|
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
if (buffer)
|
|
|
|
return buffer;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* vmalloc failed, lets dig into swap here
|
|
|
|
*/
|
|
|
|
gfp_flags &= ~__GFP_NORETRY;
|
|
|
|
buffer = (char *)__get_free_pages(gfp_flags, order);
|
|
|
|
if (buffer)
|
|
|
|
return buffer;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* complete and utter failure
|
|
|
|
*/
|
|
|
|
return NULL;
|
2005-12-06 17:38:35 -07:00
|
|
|
}
|
|
|
|
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
static struct pgv *alloc_pg_vec(struct tpacket_req *req, int order)
|
2005-12-06 17:38:35 -07:00
|
|
|
{
|
|
|
|
unsigned int block_nr = req->tp_block_nr;
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
struct pgv *pg_vec;
|
2005-12-06 17:38:35 -07:00
|
|
|
int i;
|
|
|
|
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
pg_vec = kcalloc(block_nr, sizeof(struct pgv), GFP_KERNEL);
|
2005-12-06 17:38:35 -07:00
|
|
|
if (unlikely(!pg_vec))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
for (i = 0; i < block_nr; i++) {
|
2010-11-30 19:52:57 -07:00
|
|
|
pg_vec[i].buffer = alloc_one_pg_vec_page(order);
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
if (unlikely(!pg_vec[i].buffer))
|
2005-12-06 17:38:35 -07:00
|
|
|
goto out_free_pgvec;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
return pg_vec;
|
|
|
|
|
|
|
|
out_free_pgvec:
|
|
|
|
free_pg_vec(pg_vec, order, block_nr);
|
|
|
|
pg_vec = NULL;
|
|
|
|
goto out;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2011-08-19 04:18:16 -06:00
|
|
|
static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u,
|
2009-05-18 23:11:22 -06:00
|
|
|
int closing, int tx_ring)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
struct pgv *pg_vec = NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2006-11-08 01:26:29 -07:00
|
|
|
int was_running, order = 0;
|
2009-05-18 23:11:22 -06:00
|
|
|
struct packet_ring_buffer *rb;
|
|
|
|
struct sk_buff_head *rb_queue;
|
2006-11-08 01:26:29 -07:00
|
|
|
__be16 num;
|
2011-08-19 04:18:16 -06:00
|
|
|
int err = -EINVAL;
|
|
|
|
/* Added to avoid minimal code churn */
|
|
|
|
struct tpacket_req *req = &req_u->req;
|
|
|
|
|
|
|
|
/* Opening a Tx-ring is NOT supported in TPACKET_V3 */
|
|
|
|
if (!closing && tx_ring && (po->tp_version > TPACKET_V2)) {
|
|
|
|
WARN(1, "Tx-ring is not supported.\n");
|
|
|
|
goto out;
|
|
|
|
}
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
rb = tx_ring ? &po->tx_ring : &po->rx_ring;
|
|
|
|
rb_queue = tx_ring ? &sk->sk_write_queue : &sk->sk_receive_queue;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
err = -EBUSY;
|
|
|
|
if (!closing) {
|
|
|
|
if (atomic_read(&po->mapped))
|
|
|
|
goto out;
|
|
|
|
if (atomic_read(&rb->pending))
|
|
|
|
goto out;
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
if (req->tp_block_nr) {
|
|
|
|
/* Sanity tests and some calculations */
|
|
|
|
err = -EBUSY;
|
|
|
|
if (unlikely(rb->pg_vec))
|
|
|
|
goto out;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2008-07-14 23:50:15 -06:00
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V1:
|
|
|
|
po->tp_hdrlen = TPACKET_HDRLEN;
|
|
|
|
break;
|
|
|
|
case TPACKET_V2:
|
|
|
|
po->tp_hdrlen = TPACKET2_HDRLEN;
|
|
|
|
break;
|
2011-08-19 04:18:16 -06:00
|
|
|
case TPACKET_V3:
|
|
|
|
po->tp_hdrlen = TPACKET3_HDRLEN;
|
|
|
|
break;
|
2008-07-14 23:50:15 -06:00
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
err = -EINVAL;
|
2005-12-06 17:38:35 -07:00
|
|
|
if (unlikely((int)req->tp_block_size <= 0))
|
2009-05-18 23:11:22 -06:00
|
|
|
goto out;
|
2005-12-06 17:38:35 -07:00
|
|
|
if (unlikely(req->tp_block_size & (PAGE_SIZE - 1)))
|
2009-05-18 23:11:22 -06:00
|
|
|
goto out;
|
2008-07-18 19:05:19 -06:00
|
|
|
if (unlikely(req->tp_frame_size < po->tp_hdrlen +
|
2009-05-18 23:11:22 -06:00
|
|
|
po->tp_reserve))
|
|
|
|
goto out;
|
2005-12-06 17:38:35 -07:00
|
|
|
if (unlikely(req->tp_frame_size & (TPACKET_ALIGNMENT - 1)))
|
2009-05-18 23:11:22 -06:00
|
|
|
goto out;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
rb->frames_per_block = req->tp_block_size/req->tp_frame_size;
|
|
|
|
if (unlikely(rb->frames_per_block <= 0))
|
|
|
|
goto out;
|
|
|
|
if (unlikely((rb->frames_per_block * req->tp_block_nr) !=
|
|
|
|
req->tp_frame_nr))
|
|
|
|
goto out;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
err = -ENOMEM;
|
2005-12-06 17:38:35 -07:00
|
|
|
order = get_order(req->tp_block_size);
|
|
|
|
pg_vec = alloc_pg_vec(req, order);
|
|
|
|
if (unlikely(!pg_vec))
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
2011-08-19 04:18:16 -06:00
|
|
|
switch (po->tp_version) {
|
|
|
|
case TPACKET_V3:
|
|
|
|
/* Transmit path is not supported. We checked
|
|
|
|
* it above but just being paranoid
|
|
|
|
*/
|
|
|
|
if (!tx_ring)
|
|
|
|
init_prb_bdqc(po, rb, pg_vec, req_u, tx_ring);
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
}
|
|
|
|
/* Done */
|
|
|
|
else {
|
|
|
|
err = -EINVAL;
|
2005-12-06 17:38:35 -07:00
|
|
|
if (unlikely(req->tp_frame_nr))
|
2009-05-18 23:11:22 -06:00
|
|
|
goto out;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
lock_sock(sk);
|
|
|
|
|
|
|
|
/* Detach socket from network */
|
|
|
|
spin_lock(&po->bind_lock);
|
|
|
|
was_running = po->running;
|
|
|
|
num = po->num;
|
|
|
|
if (was_running) {
|
|
|
|
po->num = 0;
|
2011-07-04 02:44:29 -06:00
|
|
|
__unregister_prot_hook(sk, false);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
spin_unlock(&po->bind_lock);
|
2007-02-09 07:25:10 -07:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
synchronize_net();
|
|
|
|
|
|
|
|
err = -EBUSY;
|
2009-01-30 15:12:06 -07:00
|
|
|
mutex_lock(&po->pg_vec_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (closing || atomic_read(&po->mapped) == 0) {
|
|
|
|
err = 0;
|
2009-05-18 23:11:22 -06:00
|
|
|
spin_lock_bh(&rb_queue->lock);
|
2010-12-10 17:02:20 -07:00
|
|
|
swap(rb->pg_vec, pg_vec);
|
2009-05-18 23:11:22 -06:00
|
|
|
rb->frame_max = (req->tp_frame_nr - 1);
|
|
|
|
rb->head = 0;
|
|
|
|
rb->frame_size = req->tp_frame_size;
|
|
|
|
spin_unlock_bh(&rb_queue->lock);
|
|
|
|
|
2010-12-10 17:02:20 -07:00
|
|
|
swap(rb->pg_vec_order, order);
|
|
|
|
swap(rb->pg_vec_len, req->tp_block_nr);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
rb->pg_vec_pages = req->tp_block_size/PAGE_SIZE;
|
|
|
|
po->prot_hook.func = (po->rx_ring.pg_vec) ?
|
|
|
|
tpacket_rcv : packet_rcv;
|
|
|
|
skb_queue_purge(rb_queue);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (atomic_read(&po->mapped))
|
2009-07-21 15:57:59 -06:00
|
|
|
pr_err("packet_mmap: vma is busy: %d\n",
|
|
|
|
atomic_read(&po->mapped));
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2009-01-30 15:12:06 -07:00
|
|
|
mutex_unlock(&po->pg_vec_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
spin_lock(&po->bind_lock);
|
2011-07-04 02:44:29 -06:00
|
|
|
if (was_running) {
|
2005-04-16 16:20:36 -06:00
|
|
|
po->num = num;
|
2011-07-04 02:44:29 -06:00
|
|
|
register_prot_hook(sk);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
spin_unlock(&po->bind_lock);
|
2011-08-19 04:18:16 -06:00
|
|
|
if (closing && (po->tp_version > TPACKET_V2)) {
|
|
|
|
/* Because we don't support block-based V3 on tx-ring */
|
|
|
|
if (!tx_ring)
|
|
|
|
prb_shutdown_retire_blk_timer(po, tx_ring, rb_queue);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
release_sock(sk);
|
|
|
|
|
|
|
|
if (pg_vec)
|
|
|
|
free_pg_vec(pg_vec, order, req->tp_block_nr);
|
|
|
|
out:
|
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2009-05-18 23:11:22 -06:00
|
|
|
static int packet_mmap(struct file *file, struct socket *sock,
|
|
|
|
struct vm_area_struct *vma)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
struct sock *sk = sock->sk;
|
|
|
|
struct packet_sock *po = pkt_sk(sk);
|
2009-05-18 23:11:22 -06:00
|
|
|
unsigned long size, expected_size;
|
|
|
|
struct packet_ring_buffer *rb;
|
2005-04-16 16:20:36 -06:00
|
|
|
unsigned long start;
|
|
|
|
int err = -EINVAL;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
if (vma->vm_pgoff)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2009-01-30 15:12:06 -07:00
|
|
|
mutex_lock(&po->pg_vec_lock);
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
expected_size = 0;
|
|
|
|
for (rb = &po->rx_ring; rb <= &po->tx_ring; rb++) {
|
|
|
|
if (rb->pg_vec) {
|
|
|
|
expected_size += rb->pg_vec_len
|
|
|
|
* rb->pg_vec_pages
|
|
|
|
* PAGE_SIZE;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (expected_size == 0)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
2009-05-18 23:11:22 -06:00
|
|
|
|
|
|
|
size = vma->vm_end - vma->vm_start;
|
|
|
|
if (size != expected_size)
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
|
|
|
|
|
|
|
start = vma->vm_start;
|
2009-05-18 23:11:22 -06:00
|
|
|
for (rb = &po->rx_ring; rb <= &po->tx_ring; rb++) {
|
|
|
|
if (rb->pg_vec == NULL)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
for (i = 0; i < rb->pg_vec_len; i++) {
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
struct page *page;
|
|
|
|
void *kaddr = rb->pg_vec[i].buffer;
|
2009-05-18 23:11:22 -06:00
|
|
|
int pg_num;
|
|
|
|
|
2010-11-30 19:52:57 -07:00
|
|
|
for (pg_num = 0; pg_num < rb->pg_vec_pages; pg_num++) {
|
|
|
|
page = pgv_to_page(kaddr);
|
2009-05-18 23:11:22 -06:00
|
|
|
err = vm_insert_page(vma, start, page);
|
|
|
|
if (unlikely(err))
|
|
|
|
goto out;
|
|
|
|
start += PAGE_SIZE;
|
packet: Enhance AF_PACKET implementation to not require high order contiguous memory allocation (v4)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Version 4 of this patch.
Change notes:
1) Removed extra memset. Didn't think kcalloc added a GFP_ZERO the way kzalloc did :)
Summary:
It was shown to me recently that systems under high load were driven very deep
into swap when tcpdump was run. The reason this happened was because the
AF_PACKET protocol has a SET_RINGBUFFER socket option that allows the user space
application to specify how many entries an AF_PACKET socket will have and how
large each entry will be. It seems the default setting for tcpdump is to set
the ring buffer to 32 entries of 64 Kb each, which implies 32 order 5
allocation. Thats difficult under good circumstances, and horrid under memory
pressure.
I thought it would be good to make that a bit more usable. I was going to do a
simple conversion of the ring buffer from contigous pages to iovecs, but
unfortunately, the metadata which AF_PACKET places in these buffers can easily
span a page boundary, and given that these buffers get mapped into user space,
and the data layout doesn't easily allow for a change to padding between frames
to avoid that, a simple iovec change is just going to break user space ABI
consistency.
So I've done this, I've added a three tiered mechanism to the af_packet set_ring
socket option. It attempts to allocate memory in the following order:
1) Using __get_free_pages with GFP_NORETRY set, so as to fail quickly without
digging into swap
2) Using vmalloc
3) Using __get_free_pages with GFP_NORETRY clear, causing us to try as hard as
needed to get the memory
The effect is that we don't disturb the system as much when we're under load,
while still being able to conduct tcpdumps effectively.
Tested successfully by me.
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Reported-by: Maciej Żenczykowski <zenczykowski@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2010-11-16 11:26:47 -07:00
|
|
|
kaddr += PAGE_SIZE;
|
2009-05-18 23:11:22 -06:00
|
|
|
}
|
2005-12-06 17:38:35 -07:00
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
2009-05-18 23:11:22 -06:00
|
|
|
|
2005-12-06 17:38:35 -07:00
|
|
|
atomic_inc(&po->mapped);
|
2005-04-16 16:20:36 -06:00
|
|
|
vma->vm_ops = &packet_mmap_ops;
|
|
|
|
err = 0;
|
|
|
|
|
|
|
|
out:
|
2009-01-30 15:12:06 -07:00
|
|
|
mutex_unlock(&po->pg_vec_lock);
|
2005-04-16 16:20:36 -06:00
|
|
|
return err;
|
|
|
|
}
|
|
|
|
|
2005-12-22 13:49:22 -07:00
|
|
|
static const struct proto_ops packet_ops_spkt = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.family = PF_PACKET,
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.release = packet_release,
|
|
|
|
.bind = packet_bind_spkt,
|
|
|
|
.connect = sock_no_connect,
|
|
|
|
.socketpair = sock_no_socketpair,
|
|
|
|
.accept = sock_no_accept,
|
|
|
|
.getname = packet_getname_spkt,
|
|
|
|
.poll = datagram_poll,
|
|
|
|
.ioctl = packet_ioctl,
|
|
|
|
.listen = sock_no_listen,
|
|
|
|
.shutdown = sock_no_shutdown,
|
|
|
|
.setsockopt = sock_no_setsockopt,
|
|
|
|
.getsockopt = sock_no_getsockopt,
|
|
|
|
.sendmsg = packet_sendmsg_spkt,
|
|
|
|
.recvmsg = packet_recvmsg,
|
|
|
|
.mmap = sock_no_mmap,
|
|
|
|
.sendpage = sock_no_sendpage,
|
|
|
|
};
|
|
|
|
|
2005-12-22 13:49:22 -07:00
|
|
|
static const struct proto_ops packet_ops = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.family = PF_PACKET,
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.release = packet_release,
|
|
|
|
.bind = packet_bind,
|
|
|
|
.connect = sock_no_connect,
|
|
|
|
.socketpair = sock_no_socketpair,
|
|
|
|
.accept = sock_no_accept,
|
2007-02-09 07:25:10 -07:00
|
|
|
.getname = packet_getname,
|
2005-04-16 16:20:36 -06:00
|
|
|
.poll = packet_poll,
|
|
|
|
.ioctl = packet_ioctl,
|
|
|
|
.listen = sock_no_listen,
|
|
|
|
.shutdown = sock_no_shutdown,
|
|
|
|
.setsockopt = packet_setsockopt,
|
|
|
|
.getsockopt = packet_getsockopt,
|
|
|
|
.sendmsg = packet_sendmsg,
|
|
|
|
.recvmsg = packet_recvmsg,
|
|
|
|
.mmap = packet_mmap,
|
|
|
|
.sendpage = sock_no_sendpage,
|
|
|
|
};
|
|
|
|
|
2009-10-04 23:58:39 -06:00
|
|
|
static const struct net_proto_family packet_family_ops = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.family = PF_PACKET,
|
|
|
|
.create = packet_create,
|
|
|
|
.owner = THIS_MODULE,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct notifier_block packet_netdev_notifier = {
|
2009-07-21 15:57:59 -06:00
|
|
|
.notifier_call = packet_notifier,
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
#ifdef CONFIG_PROC_FS
|
|
|
|
|
|
|
|
static void *packet_seq_start(struct seq_file *seq, loff_t *pos)
|
2010-02-22 00:57:18 -07:00
|
|
|
__acquires(RCU)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2007-11-19 23:31:54 -07:00
|
|
|
struct net *net = seq_file_net(seq);
|
2010-02-22 00:57:18 -07:00
|
|
|
|
|
|
|
rcu_read_lock();
|
|
|
|
return seq_hlist_start_head_rcu(&net->packet.sklist, *pos);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static void *packet_seq_next(struct seq_file *seq, void *v, loff_t *pos)
|
|
|
|
{
|
2007-12-16 15:04:02 -07:00
|
|
|
struct net *net = seq_file_net(seq);
|
2010-02-22 00:57:18 -07:00
|
|
|
return seq_hlist_next_rcu(v, &net->packet.sklist, pos);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static void packet_seq_stop(struct seq_file *seq, void *v)
|
2010-02-22 00:57:18 -07:00
|
|
|
__releases(RCU)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2010-02-22 00:57:18 -07:00
|
|
|
rcu_read_unlock();
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2007-02-09 07:25:10 -07:00
|
|
|
static int packet_seq_show(struct seq_file *seq, void *v)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
|
|
|
if (v == SEQ_START_TOKEN)
|
|
|
|
seq_puts(seq, "sk RefCnt Type Proto Iface R Rmem User Inode\n");
|
|
|
|
else {
|
2010-02-08 16:19:29 -07:00
|
|
|
struct sock *s = sk_entry(v);
|
2005-04-16 16:20:36 -06:00
|
|
|
const struct packet_sock *po = pkt_sk(s);
|
|
|
|
|
|
|
|
seq_printf(seq,
|
net: convert %p usage to %pK
The %pK format specifier is designed to hide exposed kernel pointers,
specifically via /proc interfaces. Exposing these pointers provides an
easy target for kernel write vulnerabilities, since they reveal the
locations of writable structures containing easily triggerable function
pointers. The behavior of %pK depends on the kptr_restrict sysctl.
If kptr_restrict is set to 0, no deviation from the standard %p behavior
occurs. If kptr_restrict is set to 1, the default, if the current user
(intended to be a reader via seq_printf(), etc.) does not have CAP_SYSLOG
(currently in the LSM tree), kernel pointers using %pK are printed as 0's.
If kptr_restrict is set to 2, kernel pointers using %pK are printed as
0's regardless of privileges. Replacing with 0's was chosen over the
default "(null)", which cannot be parsed by userland %p, which expects
"(nil)".
The supporting code for kptr_restrict and %pK are currently in the -mm
tree. This patch converts users of %p in net/ to %pK. Cases of printing
pointers to the syslog are not covered, since this would eliminate useful
information for postmortem debugging and the reading of the syslog is
already optionally protected by the dmesg_restrict sysctl.
Signed-off-by: Dan Rosenberg <drosenberg@vsecurity.com>
Cc: James Morris <jmorris@namei.org>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Thomas Graf <tgraf@infradead.org>
Cc: Eugene Teo <eugeneteo@kernel.org>
Cc: Kees Cook <kees.cook@canonical.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David S. Miller <davem@davemloft.net>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Eric Paris <eparis@parisplace.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2011-05-23 06:17:35 -06:00
|
|
|
"%pK %-6d %-4d %04x %-5d %1d %-6u %-6u %-6lu\n",
|
2005-04-16 16:20:36 -06:00
|
|
|
s,
|
|
|
|
atomic_read(&s->sk_refcnt),
|
|
|
|
s->sk_type,
|
|
|
|
ntohs(po->num),
|
|
|
|
po->ifindex,
|
|
|
|
po->running,
|
|
|
|
atomic_read(&s->sk_rmem_alloc),
|
|
|
|
sock_i_uid(s),
|
2009-07-21 15:57:59 -06:00
|
|
|
sock_i_ino(s));
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-07-11 00:07:31 -06:00
|
|
|
static const struct seq_operations packet_seq_ops = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.start = packet_seq_start,
|
|
|
|
.next = packet_seq_next,
|
|
|
|
.stop = packet_seq_stop,
|
|
|
|
.show = packet_seq_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int packet_seq_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
2007-11-19 23:31:54 -07:00
|
|
|
return seq_open_net(inode, file, &packet_seq_ops,
|
|
|
|
sizeof(struct seq_net_private));
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2007-02-12 01:55:36 -07:00
|
|
|
static const struct file_operations packet_seq_fops = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.owner = THIS_MODULE,
|
|
|
|
.open = packet_seq_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
2007-11-19 23:31:54 -07:00
|
|
|
.release = seq_release_net,
|
2005-04-16 16:20:36 -06:00
|
|
|
};
|
|
|
|
|
|
|
|
#endif
|
|
|
|
|
2010-01-16 20:35:32 -07:00
|
|
|
static int __net_init packet_net_init(struct net *net)
|
2007-11-19 23:28:35 -07:00
|
|
|
{
|
2010-02-22 00:57:18 -07:00
|
|
|
spin_lock_init(&net->packet.sklist_lock);
|
2007-12-11 05:19:54 -07:00
|
|
|
INIT_HLIST_HEAD(&net->packet.sklist);
|
2007-11-19 23:28:35 -07:00
|
|
|
|
|
|
|
if (!proc_net_fops_create(net, "packet", 0, &packet_seq_fops))
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2010-01-16 20:35:32 -07:00
|
|
|
static void __net_exit packet_net_exit(struct net *net)
|
2007-11-19 23:28:35 -07:00
|
|
|
{
|
|
|
|
proc_net_remove(net, "packet");
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct pernet_operations packet_net_ops = {
|
|
|
|
.init = packet_net_init,
|
|
|
|
.exit = packet_net_exit,
|
|
|
|
};
|
|
|
|
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
static void __exit packet_exit(void)
|
|
|
|
{
|
|
|
|
unregister_netdevice_notifier(&packet_netdev_notifier);
|
2007-11-19 23:28:35 -07:00
|
|
|
unregister_pernet_subsys(&packet_net_ops);
|
2005-04-16 16:20:36 -06:00
|
|
|
sock_unregister(PF_PACKET);
|
|
|
|
proto_unregister(&packet_proto);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __init packet_init(void)
|
|
|
|
{
|
|
|
|
int rc = proto_register(&packet_proto, 0);
|
|
|
|
|
|
|
|
if (rc != 0)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
sock_register(&packet_family_ops);
|
2007-11-19 23:28:35 -07:00
|
|
|
register_pernet_subsys(&packet_net_ops);
|
2005-04-16 16:20:36 -06:00
|
|
|
register_netdevice_notifier(&packet_netdev_notifier);
|
|
|
|
out:
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
module_init(packet_init);
|
|
|
|
module_exit(packet_exit);
|
|
|
|
MODULE_LICENSE("GPL");
|
|
|
|
MODULE_ALIAS_NETPROTO(PF_PACKET);
|