generic-library/openssl
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
7,473 Commits 16 Branches 258 Tags 86 MiB
c09a0318b7
Commit Graph

89 Commits

Author SHA1 Message Date
Andy Polyakov
c09a0318b7 Minor PPC assembler updates. 2006-05-03 14:07:34 +00:00
Andy Polyakov
2c5d4daac5 Yet another "teaser" Montgomery multiplication module, for PowerPC. 2006-04-30 21:15:29 +00:00
Andy Polyakov
7a5dbeb782 Minor sparcv9 clean-ups. 2005-12-27 21:27:39 +00:00
Andy Polyakov
3b4a0225e2 As SPARCV9 CPU flavor is [expected to be] detected at run-time, we can 
afford to relax SPARCV9/8+ compiler command line and produce "unversal"
binaries as we used to.
 2005-12-19 09:10:06 +00:00
Andy Polyakov
a00e414faf Unify sparcv9 assembler naming and build rules among 32- and 64-bit builds. 
Engage run-time switch between bn_mul_mont_fpu and bn_mul_mont_int.
 2005-12-16 17:39:57 +00:00
Andy Polyakov
68ea60683a Add IALU-only bn_mul_mont for SPARCv9. See commentary section for details. 2005-12-15 22:43:33 +00:00
Andy Polyakov
6df8c74d5b Switch 64-bit sparcv9 platforms from bn(64,64) to bn(64,32). This doesn't 
have impact on performance, because amount of multiplications does not
increase with this switch, not on sparcv9 that is. On the contrary, it
actually improves performance, because it spares a load of instructions
used to chase carries. Not to mention that BN assembler modules can be
shared more freely between 32- and 64-bit builts.
 2005-12-15 22:40:58 +00:00
Andy Polyakov
07645deeb8 Apply "better safe than sorry" approach after addressing sporadic SEGV in 
bn_sub_words to the rest of the sparcv8plus.S.
 2005-11-15 08:02:10 +00:00
Andy Polyakov
c52c82ffc1 Attempt to resolve sporadic SEGV crashes in bn_sub_words in OpenSSH. I'm 
baffled why it crashes and does it sporadically...
 2005-11-11 20:07:07 +00:00
Andy Polyakov
a4d729f31d Clarify binary compatibility with HAL/Fujitsu SPARC64 family. 2005-10-25 15:39:47 +00:00
Andy Polyakov
aa2be094ae Add support for 32-bit ABI to sparcv9a-mont.pl module. 2005-10-22 18:16:09 +00:00
Andy Polyakov
4d524040bc Change bn_mul_mont declaration and BN_MONT_CTX. Update CHANGES. 2005-10-22 17:57:18 +00:00
Andy Polyakov
bcb43bb358 Yet another "teaser" Montgomery multiply module, for UltraSPARC. It's not 
integrated yet, but it's tested and benchmarked [see commentary section
for further details].
 2005-10-19 07:12:06 +00:00
Andy Polyakov
34736de4c0 Flip saved argument block and tp [required for non-SSE2 path]. 2005-10-14 16:05:21 +00:00
Andy Polyakov
5f50d597f2 Make sure x86-mont.pl returns zero even if compiled with no-sse2. 2005-10-14 15:24:06 +00:00
Andy Polyakov
35593b33f4 Add timestamp to x86-mont.pl. 2005-10-09 10:26:56 +00:00
Andy Polyakov
54f3d200d3 Throw in bn/asm/x86-mont.pl Montgomery multiplication "teaser". 2005-10-09 09:53:58 +00:00
Andy Polyakov
7a2f4cbfe8 x86_64-mont.pl readability improvement. 2005-10-07 15:18:16 +00:00
Andy Polyakov
5ac7bde7c9 Throw in Montgomery multiplication assembler for x86_64. 2005-10-07 14:18:06 +00:00
Andy Polyakov
6f9afa68cd IA-32 BN tune-up. Performance imrpovement varies with platform and 
keylength, this time larger improvement for shorter keys, and reaches
15%. Both SSE2 and IALU code pathes are improved.
 2005-09-20 12:26:54 +00:00
Andy Polyakov
7cfe2a5e65 Fix Intel assembler warnings. 2005-08-10 08:28:36 +00:00
Andy Polyakov
ef428d5681 Fix unwind directives in IA-64 assembler modules. This helps symbolic 
debugging and doesn't affect functionality.

Submitted by: David Mosberger

Obtained from: http://www.hpl.hp.com/research/linux/crypto/
 2005-07-18 09:54:14 +00:00
Andy Polyakov
31efffbdba Trap condition should be 64-bit when it's due. 2005-07-03 09:17:50 +00:00
Andy Polyakov
aaa5dc614f More elegant solution to "sparse decimal printout on PPC" problem. 2005-07-02 08:58:55 +00:00
Andy Polyakov
8be97c01d1 Decimal printout of a BN is wrong on PPC, it's sparse with very few 
significant digits. As soon it verifies elsewhere it goes to 0.9.8 and
0.9.7.
 2005-07-01 17:49:47 +00:00
Richard Levitte
4bb61becbb Add emacs cache files to .cvsignore. 2005-04-11 14:17:07 +00:00
Andy Polyakov
60fd574cdf Make bn/asm/x86_64-gcc.c gcc4 savvy. +r is likely to be initially 
introduced for a reason [like bug in initial gcc port], but proposed
=&r is treated correctly by senior 3.2, so we can assume it's safe now.
PR: 1031
 2005-04-03 18:53:29 +00:00
Andy Polyakov
da30c74a27 Remove unused assembler modules. 2005-02-06 13:43:02 +00:00
Andy Polyakov
2b247cf81f OPENSSL_ia32cap final touches. Note that OPENSSL_ia32cap is no longer a 
symbol, but a macro expanded as (*(OPENSSL_ia32cap_loc())). The latter
is the only one to be exported to application.
 2004-08-29 16:36:05 +00:00
Andy Polyakov
859ceeeb51 Anchor AES and SHA-256/-512 assembler from C. 2004-07-18 17:26:01 +00:00
Andy Polyakov
e2f2a9af2c New scalable bn_mul_add_words loop, which provides up to >20% overall 
performance improvement. Make module more gcc friendly and clarify
copyright issues for division routine.
 2004-07-01 11:10:38 +00:00
Andy Polyakov
1809e858bb Eliminate compiler warnings and throw in performance table. 2004-05-28 10:15:58 +00:00
Andy Polyakov
d3adc3d3ed SSE2 accelerated bn_mul_add_words. Code is currently disabled till proper 
config and run-time support is added.
PR: 788
Submitted by: <dean@arctic.org>
Reviewed by: <appro>

Obtained from: http://arctic.org/~dean/crypto/rsa.html
 2004-05-06 10:36:49 +00:00
Andy Polyakov
dd55880644 Improved PowerPC support. Proper ./config support for ppc targets, 
especially for AIX. But most important BIGNUM assembler implementation
submitted by IBM.

Submitted by: Peter Waltenberg <pwalten@au1.ibm.com>
Reviewed by: appro
 2004-04-27 22:05:50 +00:00
Andy Polyakov
1751034669 Typo in crypto/bn/asm/x86_64.c, bn_div_words(). 
PR: 821
 2004-02-07 09:51:28 +00:00
Andy Polyakov
722d17cbac This is an *initial* tune-up. This update puts Itanium2 back on par with 
Itanium. I mean if overall performance improvement over C version was X
for Itanium, it's X even for Itanium2.
 2003-01-19 21:29:59 +00:00
Richard Levitte
2f09524501 A few more files to ignore 2003-01-16 21:32:56 +00:00
Andy Polyakov
46a0d4fbcb Support for ILP32 on HPUX-IA64. 2003-01-03 15:10:46 +00:00
Andy Polyakov
04945fda66 pa-risc2.s was not PIC, see RT#426. I strip call to fprintf as it's 
never called anyway (it's a debugging assertion). If pa-risc2W.s is
PIC remains to be seen...
 2003-01-03 10:52:40 +00:00
Richard Levitte
e9883d285d Finally, a bn_div_words() in VAX assembler that goes through all tests. 
PR: 413
 2002-12-23 11:25:51 +00:00
Richard Levitte
9b58214e4a More accurate comments. 2002-12-20 16:38:36 +00:00
Andy Polyakov
2f98abbcb6 x86_64 performance patch. 2002-12-14 20:42:05 +00:00
Richard Levitte
6ab285bf4c I think I got it now. Apparently, the case of having to shift down 
the divisor was a bit more complex than I first saw.  The lost bit
can't just be discarded, as there are cases where it is important.
For example, look at dividing 320000 with 80000 vs. 80001 (all
decimals), the difference is crucial.  The trick here is to check if
that lost bit was 1, and in that case, do the following:

1. subtract the quotient from the remainder
2. as long as the remainder is negative, add the divisor (the whole
   divisor, not the shofted down copy) to it, and decrease the
   quotient by one.

There's probably a nice mathematical proof for this already, but I
won't bother with that, unless someone requests it from me.
 2002-12-02 21:31:45 +00:00
Richard Levitte
1d3159bcca Make some names consistent. 2002-12-02 02:40:27 +00:00
Richard Levitte
f60ceb54eb Through some experimentation and thinking, I think I finally got the 
proper implementation of bn_div_words() for VAX.

If the tests go through well, the next step will be to test on Alpha.
 2002-12-02 02:28:27 +00:00
Richard Levitte
0f995b2f40 Small bugfix: even when r == d, we need to adjust r and q. 
PR: 366
 2002-12-01 02:17:23 +00:00
Richard Levitte
a678430602 Redo the VAX assembler version of bn_div_words(). 
PR: 366
 2002-12-01 00:49:36 +00:00
Andy Polyakov
622d3d3592 Support for Intel and HP-UXi assemblers. 2001-07-30 15:54:13 +00:00
Andy Polyakov
19a6e8b32c This fixes "Spurious test failures on IRIX?" reported in April. Apparently 
I was wrong in conclusions about when addition starts overflowing in combaX
routines.
 2001-06-22 19:17:42 +00:00
Andy Polyakov
52c0d30078 Get rid of "possible WAW dependency" warnings. 
Submitted by:
Reviewed by:
PR:
 2001-06-11 12:47:52 +00:00