kernel-ark/crypto
Mathias Krause 66be895158 crypto: sha1 - SSSE3 based SHA1 implementation for x86-64
This is an assembler implementation of the SHA1 algorithm using the
Supplemental SSE3 (SSSE3) instructions or, when available, the
Advanced Vector Extensions (AVX).

Testing with the tcrypt module shows the raw hash performance is up to
2.3 times faster than the C implementation, using 8k data blocks on a
Core 2 Duo T5500. For the smalest data set (16 byte) it is still 25%
faster.

Since this implementation uses SSE/YMM registers it cannot safely be
used in every situation, e.g. while an IRQ interrupts a kernel thread.
The implementation falls back to the generic SHA1 variant, if using
the SSE/YMM registers is not possible.

With this algorithm I was able to increase the throughput of a single
IPsec link from 344 Mbit/s to 464 Mbit/s on a Core 2 Quad CPU using
the SSSE3 variant -- a speedup of +34.8%.

Saving and restoring SSE/YMM state might make the actual throughput
fluctuate when there are FPU intensive userland applications running.
For example, meassuring the performance using iperf2 directly on the
machine under test gives wobbling numbers because iperf2 uses the FPU
for each packet to check if the reporting interval has expired (in the
above test I got min/max/avg: 402/484/464 MBit/s).

Using this algorithm on a IPsec gateway gives much more reasonable and
stable numbers, albeit not as high as in the directly connected case.
Here is the result from an RFC 2544 test run with a EXFO Packet Blazer
FTB-8510:

 frame size    sha1-generic     sha1-ssse3    delta
    64 byte     37.5 MBit/s    37.5 MBit/s     0.0%
   128 byte     56.3 MBit/s    62.5 MBit/s   +11.0%
   256 byte     87.5 MBit/s   100.0 MBit/s   +14.3%
   512 byte    131.3 MBit/s   150.0 MBit/s   +14.2%
  1024 byte    162.5 MBit/s   193.8 MBit/s   +19.3%
  1280 byte    175.0 MBit/s   212.5 MBit/s   +21.4%
  1420 byte    175.0 MBit/s   218.7 MBit/s   +25.0%
  1518 byte    150.0 MBit/s   181.2 MBit/s   +20.8%

The throughput for the largest frame size is lower than for the
previous size because the IP packets need to be fragmented in this
case to make there way through the IPsec tunnel.

Signed-off-by: Mathias Krause <minipli@googlemail.com>
Cc: Maxim Locktyukhin <maxim.locktyukhin@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2011-08-10 19:00:29 +08:00
..
async_tx net: remove mm.h inclusion from netdevice.h 2011-06-21 19:17:20 -07:00
ablkcipher.c crypto: skcipher - remove redundant NULL check 2011-01-29 15:09:43 +11:00
aead.c
aes_generic.c
af_alg.c atomic: use <linux/atomic.h> 2011-07-26 16:49:47 -07:00
ahash.c crypto: hash - Fix handling of small unaligned buffers 2010-08-06 09:26:38 +08:00
algapi.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6 2010-05-03 11:28:58 +08:00
algboss.c crypto: testmgr - Fix test disabling option 2010-08-06 09:40:28 +08:00
algif_hash.c crypto: algif_hash - Handle initial af_alg_make_sg error correctly 2011-06-30 07:44:06 +08:00
algif_skcipher.c crypto: algif_skcipher - Handle unaligned receive buffer 2010-11-30 17:04:31 +08:00
ansi_cprng.c Fix common misspellings 2011-03-31 11:26:23 -03:00
anubis.c
api.c
arc4.c crypto: arc4 - Fixed coding style issues 2011-06-30 07:44:05 +08:00
authenc.c crypto: Use scatterwalk_crypto_chain 2010-12-02 14:47:16 +08:00
authencesn.c crypto: authencesn - Add algorithm to handle IPsec extended sequence numbers 2011-03-13 20:22:27 -07:00
blkcipher.c mm: strictly nested kmap_atomic() 2010-10-26 16:52:08 -07:00
blowfish.c
camellia.c
cast5.c crypto: cast5 - simplify if-statements 2010-11-13 21:47:55 +09:00
cast6.c
cbc.c
ccm.c
chainiv.c
cipher.c
compress.c
crc32c.c crypto: crc32c - Fixed coding style issue 2011-06-30 07:44:05 +08:00
cryptd.c crypto: cryptd - Adding the AEAD interface type support to cryptd 2010-09-20 16:05:12 +08:00
crypto_null.c
crypto_wq.c crypto: mark crypto workqueues CPU_INTENSIVE 2011-01-04 23:34:08 +11:00
ctr.c crypto: Use ERR_CAST 2010-05-26 10:36:51 +10:00
cts.c
deflate.c net+crypto: Use vmalloc for zlib inflate buffers. 2011-06-29 05:48:41 -07:00
des_generic.c Blackfin: Rename DES PC2() symbol to avoid collision 2010-10-07 14:08:50 +01:00
ecb.c
eseqiv.c crypto: Use scatterwalk_crypto_chain 2010-12-02 14:47:16 +08:00
fcrypt.c
fips.c
gcm.c crypto: Use scatterwalk_crypto_chain 2010-12-02 14:47:16 +08:00
gf128mul.c crypto: gf128mul - fix call to memset() 2011-07-08 17:21:21 +08:00
ghash-generic.c
hmac.c
internal.h
Kconfig crypto: sha1 - SSSE3 based SHA1 implementation for x86-64 2011-08-10 19:00:29 +08:00
khazad.c
krng.c
lrw.c
lzo.c
Makefile crypto: authencesn - Add algorithm to handle IPsec extended sequence numbers 2011-03-13 20:22:27 -07:00
md4.c
md5.c
michael_mic.c
pcbc.c
pcompress.c
pcrypt.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6 2011-01-13 10:25:58 -08:00
proc.c atomic: use <linux/atomic.h> 2011-07-26 16:49:47 -07:00
ripemd.h
rmd128.c crypto: ripemd - Set module author and update email address 2011-01-04 23:34:03 +11:00
rmd160.c crypto: ripemd - Set module author and update email address 2011-01-04 23:34:03 +11:00
rmd256.c crypto: ripemd - Set module author and update email address 2011-01-04 23:34:03 +11:00
rmd320.c crypto: ripemd - Set module author and update email address 2011-01-04 23:34:03 +11:00
rng.c atomic: use <linux/atomic.h> 2011-07-26 16:49:47 -07:00
salsa20_generic.c
scatterwalk.c crypto: scatterwalk - Fix scatterwalk_done() test 2010-05-19 14:06:29 +10:00
seed.c
seqiv.c
serpent.c
sha1_generic.c crypto: sha1 - export sha1_update for reuse 2011-08-10 19:00:28 +08:00
sha256_generic.c
sha512_generic.c
shash.c crypto: hash - Fix async import on shash algorithm 2010-11-04 14:48:37 -04:00
tcrypt.c crypto: tcrypt - CTR mode speed test for AES 2011-05-04 15:06:37 +10:00
tcrypt.h
tea.c
testmgr.c crypto: testmgr - add support for aes ofb mode 2011-05-04 15:04:10 +10:00
testmgr.h crypto: testmgr - add xts-aes-256 self-test 2011-06-30 07:44:00 +08:00
tgr192.c
twofish_common.c
twofish_generic.c crypto: twofish: Rename twofish to twofish_generic and add an alias 2010-06-03 21:02:51 +10:00
vmac.c Fix common misspellings 2011-03-31 11:26:23 -03:00
wp512.c
xcbc.c
xor.c
xts.c Fix common misspellings 2011-03-31 11:26:23 -03:00
zlib.c net+crypto: Use vmalloc for zlib inflate buffers. 2011-06-29 05:48:41 -07:00