Commit graph

322 commits

Author SHA1 Message Date
Andy Polyakov
8fc8f486f7 crypto/bn/x86_64-mont5.pl: constant-time gather procedure.
At the same time remove miniscule bias in final subtraction.
Performance penalty varies from platform to platform, and even with
key length. For rsa2048 sign it was observed to be 4% for Sandy
Bridge and 7% on Broadwell.

CVE-2016-0702

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-01 13:34:22 +00:00
Andy Polyakov
d6d422e1ec bn/asm/rsaz-avx2.pl: constant-time gather procedure.
Performance penalty is 2%.

CVE-2016-0702

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-01 13:34:22 +00:00
Andy Polyakov
5ea08bd2fe bn/asm/rsax-x86_64.pl: constant-time gather procedure.
Performance penalty is 2% on Linux and 5% on Windows.

CVE-2016-0702

Reviewed-by: Richard Levitte <levitte@openssl.org>
Reviewed-by: Rich Salz <rsalz@openssl.org>
2016-03-01 13:34:22 +00:00
FdaSilvaYY
0d4fb84390 GH601: Various spelling fixes.
Signed-off-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
2016-02-05 15:25:50 -05:00
Andy Polyakov
b859d70d4a bn/asm/bn-c64xplus.asm: update commentary.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-12-22 10:30:03 +01:00
Andy Polyakov
b974943234 x86_64 assembly pack: tune clang version detection even further.
RT#4171

Reviewed-by: Kurt Roeckx <kurt@openssl.org>
2015-12-13 22:18:18 +01:00
Andy Polyakov
a285992763 ARMv4 assembly pack: allow Thumb2 even in iOS build,
and engage it in most modules.

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-12-07 12:06:06 +01:00
Andy Polyakov
29851264f1 bn/asm/x86_64-mont5.pl: fix carry propagating bug (CVE-2015-3193).
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-12-03 13:12:01 +00:00
Andy Polyakov
76eba0d94b x86_64 assembly pack: tune clang version detection.
RT#4142

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-11-23 16:00:06 +01:00
Andy Polyakov
e4693b4e2a bn/asm/ppc64-mont.pl: adapt for little-endian.
The problem remained unnoticed so far, because it's never called by default.
You have to craft OPENSSL_ppccap environment variable to trigger the problem.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-11-16 13:40:41 +01:00
Andy Polyakov
9d0e4dc635 bn/asm/s390x.S: improve performance on z196 and z13 by up to 26%. [even z10 is couple percent faster]. Triggered by RT#4128, but solves the problem by real modulo-scheduling.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-11-16 13:10:45 +01:00
Alessandro Ghedini
d900a015b5 Fix typos
Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-10-23 19:52:08 +02:00
Andy Polyakov
8eed3289b2 bn/asm/armv4-mont.pl: boost NEON performance.
Close difference gap on Cortex-A9, which resulted in further improvement
even on other processors.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-09-30 17:48:30 +02:00
Andy Polyakov
11208dcfb9 ARMv4 assembly pack: implement support for Thumb2.
As some of ARM processors, more specifically Cortex-Mx series, are
Thumb2-only, we need to support Thumb2-only builds even in assembly.

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-09-25 13:34:02 +02:00
Richard Levitte
053fa39af6 Conversion to UTF-8 where needed
This leaves behind files with names ending with '.iso-8859-1'.  These
should be safe to remove.  If something went wrong when re-encoding,
there will be some files with names ending with '.utf8' left behind.

Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-07-14 01:10:01 +02:00
Andy Polyakov
69567687b0 bn/asm/x86_64-mont5.pl: fix valgrind error.
bn_get_bits5 was overstepping array boundary by 1 byte. It was exclusively
read overstep and data could not have been used. The only potential problem
would be if array happens to end on the very edge of last accesible page.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-05-24 21:30:39 +02:00
Andy Polyakov
579734ced6 bn/asm/vis3-mont.pl: fix intermittent EC failures on SPARC T3.
BLKINIT optimization worked on T4, but for some reason appears "too
aggressive" for T3 triggering intermiitent EC failures. It's not clear
why only EC is affected...

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-05-20 09:11:25 +02:00
Andy Polyakov
d38f1b39f1 bn/asm/armv8-mont.pl: boost performance.
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-05-13 17:14:00 +02:00
Andy Polyakov
313e6ec11f Add assembly support for 32-bit iOS.
Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-04-20 15:06:22 +02:00
Andy Polyakov
cb2ed54582 Add ARMv8 Montgomery multiplication module.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-04-20 14:39:34 +02:00
Rich Salz
c436e05bdc Remove unused eng_rsax and related asm file
Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-24 16:27:03 -05:00
Matt Caswell
0f113f3ee4 Run util/openssl-format-source -v -c .
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-22 09:20:09 +00:00
Andy Polyakov
c27310f938 bn/asm/x86_64-gcc.cL make it indent-friendly.
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-22 09:20:08 +00:00
Andy Polyakov
b3d7294976 Add Broadwell performance results.
Reviewed-by: Emilia Käsper <emilia@openssl.org>
2015-01-13 21:40:14 +01:00
Andy Polyakov
a7a44ba55c Fix for CVE-2014-3570 (with minor bn_asm.c revamp).
Reviewed-by: Emilia Kasper <emilia@openssl.org>
2015-01-08 15:49:45 +00:00
Andy Polyakov
c1669e1c20 Remove inconsistency in ARM support.
This facilitates "universal" builds, ones that target multiple
architectures, e.g. ARMv5 through ARMv7. See commentary in
Configure for details.

Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-04 23:45:08 +01:00
Tim Hudson
1d97c84351 mark all block comments that need format preserving so that
indent will not alter them when reformatting comments

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Matt Caswell <matt@openssl.org>
2014-12-30 22:10:26 +00:00
Rich Salz
8cfe08b4ec Remove all .cvsignore files
Reviewed-by: Tim Hudson <tjh@openssl.org>
2014-11-28 18:32:43 -05:00
Rich Salz
df8c39d522 RT3549: Remove obsolete files in crypto
Reviewed-by: Andy Polyakov <appro@openssl.org>
2014-10-01 16:05:47 -04:00
Andy Polyakov
569e2d1257 crypto/bn/asm/x86_64-mont*.pl: add missing clang detection.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2014-09-12 00:44:51 +02:00
Andy Polyakov
15735e4f0e bn/asm/rsaz-*.pl: allow spaces in Perl path name.
RT: 2835

Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
2014-08-21 00:17:45 +02:00
Andy Polyakov
1b0fe79f3e x86_64 assembly pack: improve masm support. 2014-07-09 20:08:01 +02:00
Andy Polyakov
a356e488ad x86_64 assembly pack: refine clang detection. 2014-06-28 17:23:21 +02:00
Andy Polyakov
406d4af050 bn/asm/rsaz-avx2.pl: fix occasional failures. 2014-06-27 22:41:58 +02:00
Andy Polyakov
f3f620e1e0 bn_exp.c: move check for AD*X to rsaz-avx2.pl.
This ensures high performance is situations when assembler supports
AVX2, but not AD*X.
2014-06-27 00:07:15 +02:00
Andy Polyakov
7eb0488280 x86_64 assembly pack: addendum to last clang commit. 2014-06-24 08:37:05 +02:00
Andy Polyakov
ac171925ab x86_64 assembly pack: allow clang to compile AVX code. 2014-06-24 08:24:25 +02:00
Andy Polyakov
5dcf70a1c5 ARM assembly pack: get ARMv7 instruction endianness right.
Pointer out and suggested by: Ard Biesheuvel.
2014-06-06 21:27:18 +02:00
Andy Polyakov
bd227733b9 C64x+ assembly pack: make it work with older toolchain. 2014-05-04 16:38:32 +02:00
Andy Polyakov
f8cee9d081 bn/asm/armv4-gf2m.pl, modes/asm/ghash-armv4.pl: faster multiplication
algorithm suggested in following paper:

Câmara, D.; Gouvêa, C. P. L.; López, J. & Dahab, R.: Fast Software
Polynomial Multiplication on ARM Processors using the NEON Engine.

http://conradoplg.cryptoland.net/files/2010/12/mocrysen13.pdf
2014-04-24 10:24:53 +02:00
Andy Polyakov
eedab5241e bn/asm/x86_64-mont5.pl: fix compilation error on Solaris. 2014-01-09 13:44:59 +01:00
Andy Polyakov
ec9cc70f72 bn/asm/x86_64-mont5.pl: add MULX/AD*X code path.
This also eliminates code duplication between x86_64-mont and x86_64-mont
and optimizes even original non-MULX code.
2013-12-09 21:02:24 +01:00
Andy Polyakov
d1671f4f1a bn/asm/armv4-mont.pl: add NEON code path. 2013-12-04 22:37:49 +01:00
Andy Polyakov
c5d5f5bd0f bn/asm/x86_64-mont5.pl: comply with Win64 ABI.
PR: 3189
Submitted by: Oscar Ciurana
2013-12-03 23:59:55 +01:00
Andy Polyakov
8bd7ca9996 crypto/bn/asm/rsaz-x86_64.pl: make it work on Win64. 2013-12-03 22:28:48 +01:00
Andy Polyakov
31ed9a2131 crypto/bn/rsaz*: fix licensing note.
rsaz_exp.c: harmonize line terminating;
asm/rsaz-*.pl: minor optimizations.
2013-12-03 22:08:29 +01:00
Andy Polyakov
6efef384c6 bn/asm/rsaz-x86_64.pl: fix prototype. 2013-12-03 09:43:06 +01:00
Andy Polyakov
b9e87d07cb ppc64-mont.pl: eliminate dependency on GPRs' upper halves. 2013-11-27 22:50:00 +01:00
Andy Polyakov
4eeb750d20 bn/asm/x86_64-mont.pl: minor optimization [for Decoded ICache]. 2013-10-25 10:14:20 +02:00
Andy Polyakov
d6019e1654 PPC assembly pack: add .size directives. 2013-10-15 00:14:39 +02:00
Andy Polyakov
30b9c2348d bn/asm/*x86_64*.pl: correct assembler requirement for ad*x. 2013-10-14 22:41:00 +02:00
Andy Polyakov
0c2adb0a9b MIPS assembly pack: get rid of deprecated instructions.
Latest MIPS ISA specification declared 'branch likely' instructions
obsolete. To makes code future-proof replace them with equivalent.
2013-10-13 13:14:52 +02:00
Andy Polyakov
fa104be35e bn/asm/rsax-avx2.pl: minor optimization [for Decoded ICache]. 2013-10-10 23:06:43 +02:00
Andy Polyakov
a5bb5bca52 bn/asm/x86_64-mont*.pl: add MULX/ADCX/ADOX code path. 2013-10-03 00:45:04 +02:00
Andy Polyakov
87954638a6 rsaz-x86_64.pl: add MULX/ADCX/ADOX code path. 2013-10-03 00:30:12 +02:00
Andy Polyakov
72a158703b crypto/bn/asm/x86_64-mont.pl: minor optimization. 2013-09-09 21:40:33 +02:00
Veres Lajos
478b50cf67 misspellings fixes by https://github.com/vlajos/misspell_fixer 2013-09-05 21:39:42 +01:00
Andy Polyakov
fd8ad019e1 crypto/bn/asm/rsax-x86_64.pl: make it work on Darwin. 2013-08-03 16:28:50 +02:00
Andy Polyakov
5c57c69f9e bn/asm/rsaz-avx2.pl: Windows-specific fix. 2013-07-12 18:59:17 +02:00
Ben Laurie
852f837f5e s/rsaz_eligible/rsaz_avx2_eligible/. 2013-07-12 12:47:39 +01:00
Andy Polyakov
0b4bb91db6 Add RSAZ assembly modules.
RT: 2582, 2850
2013-07-05 21:30:18 +02:00
Andy Polyakov
26e43b48a3 bn/asm/x86_86-mont.pl: optimize reduction for Intel Core family. 2013-07-05 21:10:56 +02:00
Andy Polyakov
4ddacd9921 Optimize SPARC T4 MONTMUL support.
Improve RSA sing performance by 20-30% by:
- switching from floating-point to integer conditional moves;
- daisy-chaining sqr-sqr-sqr-sqr-sqr-mul sequences;
- using MONTMUL even during powers table setup;
2013-06-18 10:39:38 +02:00
Andy Polyakov
02450ec69d PA-RISC assembler pack: switch to bve in 64-bit builds.
PR: 3074
2013-06-18 10:37:00 +02:00
Adam Langley
7753a3a684 Add volatile qualifications to two blocks of inline asm to stop GCC from
eliminating them as dead code.

Both volatile and "memory" are used because of some concern that the compiler
may still cache values across the asm block without it, and because this was
such a painful debugging session that I wanted to ensure that it's never
repeated.
2013-06-04 18:46:25 +01:00
Andy Polyakov
342dbbbe4e x86_64-gf2m.pl: fix typo. 2013-03-01 22:36:36 +01:00
Andy Polyakov
7c43601d44 x86_64-gf2m.pl: add missing Windows build fix for #2963.
PR: 3004
2013-03-01 21:43:10 +01:00
Andy Polyakov
4568182a8b x86_64 assembly pack: keep making Windows build more robust.
PR: 2963 and a number of others
2013-02-02 19:54:59 +01:00
Andy Polyakov
46bf83f07a x86_64 assembly pack: make Windows build more robust.
PR: 2963 and a number of others
2013-01-22 22:27:28 +01:00
Andy Polyakov
543fd85460 bn/asm/mips.pl: hardwire local call to bn_div_words. 2013-01-22 21:13:37 +01:00
Andy Polyakov
904732f68b C64x+ assembly pack: improve EABI support. 2012-11-28 13:19:10 +00:00
Andy Polyakov
9f6b0635ad x86_64-gcc.c: resore early clobber constraint.
Submitted by: Florian Weimer
2012-11-19 15:02:00 +00:00
Andy Polyakov
68c06bf6b2 Support for SPARC T4 MONT[MUL|SQR] instructions.
Submitted by: David Miller, Andy Polyakov
2012-11-17 10:34:11 +00:00
Andy Polyakov
1efd583085 SPARCv9 assembly pack: harmonize ABI handling (so that it's handled in one
place at a time, by pre-processor in .S case and perl - in .s).
2012-10-25 12:07:32 +00:00
Andy Polyakov
0c832ec5c6 Add VIS3-capable sparcv9-gf2m module. 2012-10-20 15:59:14 +00:00
Andy Polyakov
947d78275b Add VIS3 Montgomery multiplication. 2012-10-20 09:13:21 +00:00
Andy Polyakov
be0d31b166 Add linux-x32 target. 2012-08-29 14:08:46 +00:00
Andy Polyakov
1a002d88ad MIPS assembly pack: assign default value to $flavour. 2012-08-17 09:10:31 +00:00
Andy Polyakov
6251989eb6 x86_64 assembly pack: make it possible to compile with Perl located on
path with spaces.

PR: 2835
2012-06-27 10:08:23 +00:00
Andy Polyakov
3e181369dd C64x+ assembler pack. linux-c64xplus build is *not* tested nor can it be
tested, because kernel is not in shape to handle it *yet*. The code is
committed mostly to stimulate the kernel development.
2012-04-18 13:01:36 +00:00
Andy Polyakov
8c98b2591f modexp512-x86_64.pl: Solaris protability fix.
PR: 2656
2011-12-12 15:10:14 +00:00
Andy Polyakov
5711dd8eac x86-mont.pl: fix bug in integer-only squaring path.
PR: 2648
2011-12-09 14:21:25 +00:00
Andy Polyakov
6600126825 bn/asm/mips.pl: fix typos. 2011-12-01 12:16:09 +00:00
Andy Polyakov
29fd6746f5 armv4cpuid.S, armv4-gf2m.pl: make newest code compilable by older assembler. 2011-11-05 13:07:18 +00:00
Andy Polyakov
09f40a3cb9 ppc.pl: fix bug in bn_mul_comba4.
PR: 2636
Submitted by: Charles Bryant
2011-11-05 10:16:04 +00:00
Andy Polyakov
a9cf0b81fa Remove superseded MIPS assembler modules. 2011-10-19 21:42:21 +00:00
Andy Polyakov
8329e2e776 bn_exp.c: further optimizations using more ideas from
http://eprint.iacr.org/2011/239.
2011-10-17 17:41:49 +00:00
Andy Polyakov
3f66f2040a x86_64-mont.pl: minor optimization. 2011-10-17 17:39:59 +00:00
Bodo Möller
cdfe0fdde6 Fix OPENSSL_BN_ASM_MONT5 for corner cases; add a test.
Submitted by: Emilia Kasper
2011-10-13 12:35:10 +00:00
Andy Polyakov
6c01cbb6a0 modexp512-x86_64.pl: make it work with ml64. 2011-08-19 06:30:32 +00:00
Andy Polyakov
e7d1363d12 x86_64-mont5.pl: add missing Win64 support. 2011-08-14 09:06:06 +00:00
Andy Polyakov
10bd69bf4f armv4-mont.pl: profiler-assisted optimization gives 8%-14% improvement
(more for longer keys) on RSA/DSA.
2011-08-13 12:38:41 +00:00
Andy Polyakov
ae8b47f07f SPARC assembler pack: fix FIPS linking errors. 2011-08-12 21:38:19 +00:00
Andy Polyakov
361512da0d This commit completes recent modular exponentiation optimizations on
x86_64 platform. It targets specifically RSA1024 sign (using ideas
from http://eprint.iacr.org/2011/239) and adds more than 10% on most
platforms. Overall performance improvement relative to 1.0.0 is ~40%
in average, with best result of 54% on Westmere. Incidentally ~40%
is average improvement even for longer key lengths.
2011-08-12 16:44:32 +00:00
Andy Polyakov
20735f4c81 alphacpuid.pl: fix alignment bug.
alpha-mont.pl: fix typo.
PR: 2577
2011-08-12 12:28:52 +00:00
Andy Polyakov
85ec54a417 x86_64-mont.pl: futher optimization resulting in up to 48% improvement
(4096-bit RSA sign benchmark on Core2) in comparison to initial version
from 2005.
2011-08-09 13:05:05 +00:00
Andy Polyakov
be9a8cc2af Add RSAX builtin engine. It optimizes RSA1024 sign benchmark. 2011-07-20 21:49:46 +00:00
Andy Polyakov
87873f4328 ARM assembler pack: add platform run-time detection. 2011-07-17 17:40:29 +00:00
Andy Polyakov
6179f06077 x86_64-mont.pl: add squaring procedure and improve RSA sign performance
by up to 38% (4096-bit benchmark on Core2).
2011-07-05 09:21:03 +00:00
Andy Polyakov
02a73e2bed s390x-gf2m.pl: commentary update (final performance numbers turned to be
higher).
2011-07-04 11:20:33 +00:00