wbrawner/openssl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
3723 commits 20 branches 302 tags 153 MiB
Commit graph

347 commits

Author SHA1 Message Date
Bodo Möller
946cd9a540 Change submitted files so that they compile (in particular, 
use BN_CTX_start/get/end instead of accessing ctx->tos).

Change indentation to "EAY" style.
 2000-11-26 19:13:52 +00:00
Bodo Möller
5acaa49504 More BN_mod_... functions. 2000-11-26 18:31:32 +00:00
Bodo Möller
535b9b5724 Add bn_mod.c (should have happend in the previous commit ...). 
BN_swap manual page.
 2000-11-26 16:46:57 +00:00
Bodo Möller
78a0c1f18d modular arithmetics 
"make update"
 2000-11-26 16:42:38 +00:00
Bodo Möller
6cc5e19d47 Remove CR at line ends. 2000-11-26 12:55:19 +00:00
Bodo Möller
7e0c5264e7 Elliptic curves over GF(p), new BIGNUM functions, Montgomery re-implementation. 
These new files will not be included literally in OpenSSL, but I intend
to integrate most of their contents.  Most file names will change,
and when the integration is done, the superfluous files will be deleted.

Submitted by: Lenka Fibikova <fibikova@exp-math.uni-essen.de>
 2000-11-26 12:12:35 +00:00
Richard Levitte
73343ac38a Correct a number of syntax errors. 2000-11-22 18:20:31 +00:00
Richard Levitte
c4f3542abf Reimplement bn_div_words, bn_add_words and bn_sub_words for VAX. 
I'm a little bit nervous about bn_div_words, as I don't know what it's
supposed to return on overflow.  For now, I trust the rest of the
system to give it numbers that will not cause any overflow...
 2000-11-21 23:32:38 +00:00
Richard Levitte
c799770060 Avoid getting warnings about unary - being used on unsigned integer. 2000-11-21 23:29:55 +00:00
Richard Levitte
baa257f1ed Remove two bn_wexpand() from BN_mul(), which is a step toward getting 
BN_mul() correctly constified, avoids two realloc()'s that aren't
really necessary and saves memory to boot.  This required a small
change in bn_mul_part_recursive() and the addition of variants of
bn_cmp_words(), bn_add_words() and bn_sub_words() that can take arrays
with differing sizes.

The test results show a performance that very closely matches the
original code from before my constification.  This may seem like a
very small win from a performance point of view, but if one remembers
that the variants of bn_cmp_words(), bn_add_words() and bn_sub_words()
are not at all optimized for the moment (and there's no corresponding
assembler code), and that their use may be just as non-optimal, I'm
pretty confident there are possibilities...

This code needs reviewing!
 2000-11-18 22:58:26 +00:00
Richard Levitte
e06cc57118 Remove a declaration for a function that does not exist. 2000-11-18 20:52:14 +00:00
Richard Levitte
0450d28759 Make the definition of bn_add_words() match the definition. 2000-11-18 20:49:02 +00:00
Richard Levitte
227385b719 Make sure BN_DIV2W is not defining when defining it, and remove the 
declarations of bn_add_part_words() and bn_sub_part_words() since they
do not exist.
 2000-11-17 12:01:55 +00:00
Bodo Möller
28fd5c60de Constify bn_dump1 implementation so that it matches the prototype 
in bn.h
 2000-11-17 08:36:10 +00:00
Richard Levitte
cbd48ba626 More constification of the BN library. 2000-11-16 22:43:32 +00:00
Richard Levitte
67d93e6f49 Make sure to print the BN counting (BN_COUNT) to stderr instead of 
stdout.  bc gets so confused by bean counts.
 2000-11-16 22:42:39 +00:00
Richard Levitte
afb83c45e6 Oops, when I clean, I should do it thoroughly. 2000-11-16 22:41:26 +00:00
Richard Levitte
43fcc1b096 I've checked again and again. There really is no need to expand a to 
4 times it's size when bn_sqr_recursive() won't look farther than the
original length.  Thereby, constification is no longer a problem.
 2000-11-16 21:35:41 +00:00
Ben Laurie
757e392d4e Make Rijndael work! Those long flights have some good points. 2000-11-12 02:13:38 +00:00
Bodo Möller
e1a8ac495b tmp2 is not used in BN_mod_mul_montgomery. 2000-11-09 16:39:01 +00:00
Bodo Möller
7f7b8d6871 BN_CTX-related fixes. 2000-11-08 10:05:34 +00:00
Bodo Möller
e0bf5c11ea Handle BN_copy failure after successful BN_new. 2000-11-07 09:39:51 +00:00
Bodo Möller
58f0f52e67 handle the case when BN_new returns NULL 2000-11-07 09:35:19 +00:00
Richard Levitte
020fc820dc Constify the BIGNUM routines a bit more. The only trouble were the 
two functions that did expansion on in parameters (BN_mul() and
BN_sqr()).  The problem was solved by making bn_dup_expand() which is
a mix of bn_expand2() and BN_dup().
 2000-11-06 21:15:54 +00:00
Richard Levitte
5270e7025e Merge the engine branch into the main trunk. All conflicts resolved. 
At the same time, add VMS support for Rijndael.
 2000-10-26 21:07:28 +00:00
Richard Levitte
6df3553255 A compiler warning removed. Thanks to the folks at HP! 2000-09-27 13:54:28 +00:00
Richard Levitte
4e20a4e688 'ranlib' doesn't always run on some systems. That's actually 
acceptable, since all that happens if it fails is a library with
an index, which makes linking slower, but still working correctly.
 2000-09-25 08:53:15 +00:00
Bodo Möller
fe03519704 Totally remove the supposedly 'faster' variant in 
BN_mod_mul_montgomery, which calls bn_sqr_recursive
without much preparation.

bn_sqr_recursive requires the length of its argument to be
a power of 2, which is not always the case here.
There's no reason for not using BN_sqr -- if a simpler
approach to squaring made sense, then why not change
BN_sqr?  (Using BN_sqr should also speed up DH where g is chosen
such that it becomes small [e.g., 2] when converted
to Montgomery representation.)

Case closed :-)
 2000-09-19 23:25:00 +00:00
Bodo Möller
a45bd29535 Document BN_mod_mul_montgomery bug; 
make disabled code slightly more correct (this does not solve
the problem though).
 2000-09-19 18:02:15 +00:00
Bodo Möller
f4364e0730 Disable buggy code variant in BN_mod_mul_montgomery that was enabled 
in 0.9.6-beta1 and 0.9.6-beta2 and caused the BN_mont_exp_mont_word()
failure (bug report "openssh 2.2.0p1 fails with openssl 0.9.6-beta1").
 2000-09-19 17:21:52 +00:00
Richard Levitte
d0c2ebf414 A patch from HP for better performance. 
Submitted by Kevin Steves <ks@hp.se> 3 months ago...
 2000-09-17 20:04:42 +00:00
Ulf Möller
1d84fd64fc Bug fix: Montgomery multiplication could produce results with the wrong 
sign.
 2000-09-14 18:37:53 +00:00
Richard Levitte
62ab514e98 'make update' 2000-09-07 08:46:51 +00:00
Bodo Möller
bbb8de0966 Avoid abort() throughout the library, except when preprocessor 
symbols for debugging are defined.
 2000-09-04 15:34:43 +00:00
Bodo Möller
5e38616380 Fix for BN_mul_word(a, 0). 2000-09-04 15:30:14 +00:00
Ben Laurie
c9b51693dc max -> dmax in bn_check_top. 2000-08-22 21:23:09 +00:00
Dr. Stephen Henson
2d978cbd30 Changes needed for Tandem NSK, supplied by Scott Uroff (scott@xypro.com). 
Fix warnings with BIO_dump_indent().
 2000-08-04 00:01:39 +00:00
Richard Levitte
8083e1bd9e There's a slight possibility that a is 0 in BN_sub_word(), and might 
therefore have unallocated parts.  Therefore, a check for the 0 case
is needed, resulting with the same thing as when a is negative.
 2000-07-27 21:32:23 +00:00
Richard Levitte
8b4e27e26e In the case where a < 0 and |a| < w, the result (assigned to a) from 
BN_add_word becomes wrongly negative...
This was discovered by Darrel Hankerson <dhankers@cacr.math.uwaterloo.ca>
 2000-07-27 21:17:14 +00:00
Richard Levitte
47ff5de8c2 Darrel Hankerson <dhankers@cacr.math.uwaterloo.ca> correctly discovered 
that of the result pointer to bn_mul was the same as one of the two arguments,
That argument might have it's sign destroyed.  He provided this fix.
 2000-07-26 01:14:16 +00:00
Bodo Möller
904cb691f3 Return bignum '0' when BN_rand is asked for a 0 bit random number. 2000-07-02 19:42:19 +00:00
Ben Laurie
69b5d3c51f Don't let top go below zero! 2000-07-01 16:30:27 +00:00
Bodo Möller
fc57ebc057 Actually comment out the parts of BN_MOD_MUL_WORD that I inteded to 
comment out in the previous commit
 2000-06-23 08:10:28 +00:00
Bodo Möller
e958c5afe7 In BN_mod_exp_mont_word, avoid one application of BN_MOD_MUL_WORD, 
and for small 'a' also a couple of calls to
BN_mod_mul_montgomery(r, r, r, ...).
 2000-06-23 05:51:37 +00:00
Ulf Möller
8415060e1e use faster version 2000-06-21 21:20:46 +00:00
Bodo Möller
24bec03beb This probably fixes a BN_rshift bug. 2000-06-13 15:07:39 +00:00
Bodo Möller
c3b1424cc0 Use BN_CTX_end when exiting early from BN_mod_exp_mont_word because 
BN_mod_exp_atalla could be used.
 2000-06-13 13:50:42 +00:00
Geoff Thorpe
8ba6d1b0cc The atalla functionality doesn't work with the "word" version of 
BN_mod_exp. Call the regular atalla mod_exp operation instead.
 2000-06-13 13:18:51 +00:00
Geoff Thorpe
1c4f90a05d Enable DSO support on alpha (OSF1), cc and gcc. 
Also, "make update" has added some missing functions to libeay.num,
updated the TABLE for the alpha changes, and updated thousands of
dependancies that have changed from recent commits.
 2000-06-13 12:59:38 +00:00
Bodo Möller
814ed26cfa Comment for increased code clarity. 2000-06-09 09:11:30 +00:00
Bodo Möller
dc434bbcb0 Slightly faster DSA verification (BN_mod_exp2_mont), 
marginally faster BN_mod_exp for 1024 bit exponents.
 2000-06-08 20:26:03 +00:00
Bodo Möller
f8989a2155 Use the equivalent of a sliding window (without precomputation 
because we're only handling words anyway) in BN_mod_exp_mont_word
making it a little faster for very small exponents,
and adjust the performance gain estimate in CHANGES according
to slightly more thorough measurements.
(15% faster than BN_mod_exp_mont for "large" base,
20% faster than BN_mod_exp_mont for small base.)
 2000-06-08 09:39:28 +00:00
Bodo Möller
6dad7bd69c Speed up DH with small generator. 2000-06-07 21:29:25 +00:00
Richard Levitte
26a3a48d65 There have been a number of complaints from a number of sources that names 
like Malloc, Realloc and especially Free conflict with already existing names
on some operating systems or other packages.  That is reason enough to change
the names of the OpenSSL memory allocation macros to something that has a
better chance of being unique, like prepending them with OPENSSL_.

This change includes all the name changes needed throughout all C files.
 2000-06-01 22:19:21 +00:00
Ulf Möller
de42b6a7a8 Use NO_FP_API. 2000-06-01 20:25:44 +00:00
Ulf Möller
623bb80b77 In some of the Makefiles CPP was not defined. 2000-03-20 18:16:52 +00:00
Bodo Möller
88f17a5e98 Remove Win32 assembler files. They are always rebuilt (with some 
choice of parameters) when they are needed.
 2000-03-13 08:04:20 +00:00
Ulf Möller
3a6a39c373 asm workaround for SuSE Linux 
proposed by Holger Reif
 2000-03-12 12:52:36 +00:00
Bodo Möller
59fc2b0fc2 Preserve reason strings in automatically build tables. 2000-03-05 00:19:36 +00:00
Bodo Möller
df83eeb710 Update comment. 2000-03-02 22:08:30 +00:00
Bodo Möller
743e7be913 Add missing dependencies. 2000-03-02 19:30:41 +00:00
Bodo Möller
2d092edf5e Change comment. 2000-03-02 14:34:08 +00:00
Ulf Möller
8782a42627 add comment. 2000-03-01 19:08:47 +00:00
Ulf Möller
0b8fa44ef0 Bug fix. 2000-03-01 19:08:27 +00:00
Ulf Möller
f1a617edc9 check for WIN32 (needed by Mingw32) 2000-03-01 01:46:34 +00:00
Richard Levitte
4ba48ec4d0 Include rand.h so RAND_pseudo_bytes may be declared. 
Remove one ampersand so the compiler may complain less.
Make rand() static so it will not conflict with the C RTL.
Make bug() static too, for good measure.
 2000-02-27 22:57:53 +00:00
Dr. Stephen Henson
587bb0e02e Don't call BN_rand with zero bits in bntest.c 2000-02-27 17:34:30 +00:00
Ben Laurie
707ef52e21 Typo. 2000-02-27 16:56:48 +00:00
Bodo Möller
981cdfab86 Add a comment. 2000-02-27 11:48:32 +00:00
Ulf Möller
688938fbb4 Bug fix! 2000-02-27 02:05:39 +00:00
Ulf Möller
1070e0e2ee *** empty log message *** 2000-02-27 00:00:56 +00:00
Ulf Möller
1b199605c4 workaround no longer needed 2000-02-26 22:37:43 +00:00
Ulf Möller
f27e41349f remove 2000-02-26 22:37:34 +00:00
Ulf Möller
775c63fc02 Reorganize bn_mul.c (no bugfix yet), remove obsolete files in BN library. 2000-02-26 22:16:47 +00:00
Andy Polyakov
63933136ab BN_div_recp fix. I've ran divtest for 10 mins and it didn't exhibit a 
single fault:-) Needless to mention that bnbug.c posted couple of days
ago passes as well...
 2000-02-26 16:20:05 +00:00
Ulf Möller
582afb4bd7 flush output. 
Looks like it fails when b is a power of 2, but I never get incorrect
results.
 2000-02-25 22:08:37 +00:00
Richard Levitte
f36e02b2ef Enhance consistency by using BIO_flush() instead of fflush(). 2000-02-25 20:40:12 +00:00
Ulf Möller
17dddc0596 Test the division functions. 
Apparently BN_div_recp reports an error for small divisors
(1,2,4,8,40).

I haven't got mismatches so far. If you can, please run the test
program for a few days (nohup divtest >out& or something), and if it
reports a mismatch, post the output.
 2000-02-25 20:28:54 +00:00
Ulf Möller
766d78c8f5 divide the correct number... 2000-02-25 00:14:50 +00:00
Ulf Möller
c04949e978 work around a bug in BN_div_recp or BN_reciprocal 2000-02-24 23:37:15 +00:00
Bodo Möller
0dd0cbf554 Do fflush(stdout) when there was an error. 2000-02-24 20:13:27 +00:00
Ulf Möller
4a26329b4c correct macro. 2000-02-23 22:11:07 +00:00
Ben Laurie
bd03b99b9b Add support for Compaq Atalla crypto accelerator. 2000-02-16 22:15:39 +00:00
Andy Polyakov
8691ff97a1 Move primes to read-only segment. 2000-02-16 13:24:06 +00:00
Andy Polyakov
89cee61b45 New NO_INLINE_ASM macro. Primary target for the moment is Solaris x86 
which can't stand GNU C assembler templates.
 2000-02-15 17:20:52 +00:00
Andy Polyakov
ea96c4bc7f test_mont was exercising 100-bit multiplication modulus X*I-bit, where 
X is 5120 on 32-bit and 151552 on 64-bit architectures and I varies
from 0 to 4. As result the test was *unreasonably* slow and virtually
impossible to complete on 64-bit architectures (e.g. IRIX bc couldn't
even swallow such long lines).
 2000-02-15 13:50:02 +00:00
Ralf S. Engelschall
667ac4ec6a Make gcc 2.95.2 happy again, even under ``-Wall -Wshadow -Wpointer-arith -Wcast-align 
-Wmissing-prototypes -Wmissing-declarations -Wnested-externs -Winline''.
 2000-02-11 09:47:18 +00:00
Dr. Stephen Henson
f07fb9b24b Add command line password options to the reamining utilities, 
amend docs.
 2000-02-08 01:34:59 +00:00
Ulf Möller
4e539aaa1e "print" is GNU bc specific. 2000-02-06 23:33:06 +00:00
Ulf Möller
fcb20be846 put missing line back in. 2000-02-06 16:35:28 +00:00
Ulf Möller
cae55bfc68 Improve bntest slightly, and fix another bug in the BN library. 2000-02-06 15:56:59 +00:00
Ulf Möller
4a6222d71b BN_div bugfix. The q-- loop should not be entered in the n0==d0 case. 2000-02-06 00:25:39 +00:00
Bodo Möller
29a28ee503 Cosmetic changes. 2000-02-05 21:28:09 +00:00
Bodo Möller
eb5a6a55c5 Commit patch to bn.h that CVS decided to throw away during 'cvs update', 
and initialize too_many because memset(..., 0, ...) is not used here.
 2000-02-05 20:39:26 +00:00
Bodo Möller
37e48b88ad Generate just one error code if iterated SSL_CTX_get() fails. 
Avoid enabled 'assert()' in production library.
 2000-02-05 19:29:00 +00:00
Ulf Möller
6535eb1728 Use MONT_WORD macro to control if the word-based or the bignum 
algorithm is used.
 2000-02-05 18:23:05 +00:00
Ulf Möller
9b141126d4 New functions BN_CTX_start(), BN_CTX_get(), BN_CTX_end() to access 
temporary BIGNUMs. BN_CTX still uses a fixed number of BIGNUMs, but
the BN_CTX implementation could now easily be changed.
 2000-02-05 14:17:32 +00:00
Ulf Möller
ce76ce43ae *** empty log message *** 2000-02-03 23:38:55 +00:00
Ulf Möller
657e60fa00 ispell (and minor modifications) 2000-02-03 23:23:24 +00:00
Bodo Möller
7999c65c9b Some 'const's for BNs. 2000-02-03 01:26:07 +00:00
Bodo Möller
bfe30e4d1b Include OpenSSL license. 2000-02-02 23:30:32 +00:00
Bodo Möller
aff0825c61 Tolerate negative numbers in BN_is_prime. 2000-02-02 22:18:01 +00:00
Bodo Möller
e74231ed9e rndsort{Miller, Rabin} primality test. 2000-02-02 21:20:44 +00:00
Andy Polyakov
fb81ac5e6b Support for "multiply high" instruction, see BN_UMULT_HIGH comment in 
crypto/bn/bn_lcl.h for further details. It should be noted that for
the moment of this writing the code was tested only on Alpha. If
compiled with DEC C the C implementation exhibits 12% performance
improvement over the crypto/bn/asm/alpha.s (on EV56 box running
AlphaLinux). GNU C is (unfortunately) 8% behind the assembler
implementation. But it's OpenVMS Alpha users who *may* benefit most
as 'apps/openssl speed rsa' exhibits 6 (six) times performance
improvement over the original VMS bignum implementation. Where "*may*"
means "as soon as code is enabled though #define SIXTY_FOUR_BIT and
crypto/bn/asm/vms.mar is skipped."
 2000-02-02 16:18:12 +00:00
Ulf Möller
1399f17a07 Bug fix: BN_is_prime() would fail with a high probability for small 
primes (negligible for larger ones).
 2000-02-01 23:48:55 +00:00
Ulf Möller
0bde1089f8 match the prototype 2000-02-01 23:47:24 +00:00
Bodo Möller
cb5b7850ac If n0 == d0, we must alway compute 'rem' "by hand" 2000-02-01 11:10:54 +00:00
Bodo Möller
fe7cd1647d Typo in preprocessor symbol. 2000-02-01 08:48:30 +00:00
Ulf Möller
f40c02d908 undo. I keep confusing my directories. :( 2000-02-01 02:24:10 +00:00
Ulf Möller
8efb60144d EBCDIC support. 
Submitted by: Martin Kraemer <martin.kraemer@mch.sni.de>
 2000-02-01 02:21:16 +00:00
Ulf Möller
51ca375e7e Seek out and destroy another evil cast. 2000-01-30 23:33:40 +00:00
Bodo Möller
1baa94907c Make output of "openssl dsaparam 1024" more interesting :-) 2000-01-30 03:32:28 +00:00
Bodo Möller
a87030a1ed Make DSA_generate_parameters, and fix a couple of bug 
(including another problem in the s3_srvr.c state machine).
 2000-01-30 02:23:03 +00:00
Richard Levitte
f95a9f678a Declare BN_pseudo_rand(). 2000-01-27 22:07:42 +00:00
Bodo Möller
953aa79040 A couple of things were reversed for BN_pseudo_rand ... 2000-01-27 21:09:25 +00:00
Ulf Möller
157be2b67d comment was wrong. 2000-01-27 19:52:58 +00:00
Ulf Möller
38e33cef15 Document DSA and SHA. 
New function BN_pseudo_rand().
Use BN_prime_checks_size(BN_num_bits(w)) rounds of Miller-Rabin when
generating DSA primes (why not use BN_is_prime()?)
 2000-01-27 19:31:26 +00:00
Bodo Möller
2233bed1cb typo in a comment 2000-01-27 14:31:31 +00:00
Bodo Möller
ec1258dd44 Update comments to provide a better approximation of reality. 2000-01-27 13:00:10 +00:00
Bodo Möller
b99b110766 Add a pointer to a paper (is the algorithm in section 4.2 the 
word-based algorithm we are using?)
 2000-01-27 12:52:41 +00:00
Bodo Möller
6e0cad8d34 enable Montgomery test 2000-01-27 11:26:28 +00:00
Bodo Möller
5f5e96d92b Update references. 2000-01-27 09:15:49 +00:00
Ulf Möller
e93f9a3284 Run ispell. 
Clean up bn_mont.c.
 2000-01-27 01:50:42 +00:00
Dr. Stephen Henson
dd9d233e2a Tidy up CRYPTO_EX_DATA structures. 2000-01-23 23:41:49 +00:00
Ulf Möller
dd8dec69b8 Document the BN library. 2000-01-23 22:06:24 +00:00
Ulf Möller
4486d0cd7a Document the DH library, and make some minor changes along the way. 2000-01-22 20:05:23 +00:00
Ulf Möller
b25c8db872 AFAICS lst1 stands for "lshift test" not "list". 2000-01-21 00:00:12 +00:00
Bodo Möller
4146aa6b42 Rename lst1 to list1 to avoid name conflict on some platforms. 2000-01-20 23:15:50 +00:00
Bodo Möller
0c50e02b30 - Pseudo-seed the PRNG in programs used for "make test" 
because otherwise BN_rand will fail unless DEVRANDOM works,
  which causes the programs to dump core because they
  don't check the return value of BN_rand (and if they
  did, we still couldn't test anything).

- add comment to some files that appear not to be used at all.
 2000-01-14 17:55:37 +00:00
Ulf Möller
eb952088f0 Precautions against using the PRNG uninitialized: RAND_bytes() now 
returns int (1 = ok, 0 = not seeded). New function RAND_add() is the
same as RAND_seed() but takes an estimate of the entropy as an additional
argument.
 2000-01-13 20:59:17 +00:00
Bodo Möller
76aa0ddc86 Turn BN_prime_checks into a macro. 
Primes p where (p-1)/2 is prime too are called "safe", not "strong".
 2000-01-12 11:57:30 +00:00
Bodo Möller
e4b7645631 add dependency and auto-generation rule for bn_prime.h 
(created by bn_prime.pl, which now prints the copyright/license
note as found in bn_prime.h)
 2000-01-12 09:16:22 +00:00
Andy Polyakov
b96eb06f79 Makefile clean-ups, crypto/bn/asm/alpha.s compiles on Alpha Linux. 1999-12-25 16:08:31 +00:00
Ulf Möller
3b14cb717d Solaris x86 assembler problem is already addressed in ./config 
(bug reports keep coming in because that was still missing in 0.9.4)
 1999-12-14 15:28:10 +00:00
Ulf Möller
a4af39ac44 Don't use inline assembler on x86 Solaris (would need a different syntax). 1999-12-08 22:55:06 +00:00
Mark J. Cox
ce2c95b2a2 Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). The 
problem was that one of the replacement routines had not been working since
SSLeay releases.  For now the offending routine has been replaced with
non-optimised assembler.  Even so, this now gives around 95% performance
improvement for 1024 bit RSA signs.
 1999-11-03 14:10:10 +00:00
Bodo Möller
798757762a Improve support for running everything as a monolithic application. 
Submitted by: Lennart Bång, Bodo Möller
 1999-10-25 19:36:01 +00:00
Ulf Möller
8e1589ece5 VC++ warning. 1999-09-29 22:11:06 +00:00
Andy Polyakov
0bbd03525e Minor MIPS III/IV tune-up. 1999-08-24 16:02:16 +00:00
Ben Laurie
97e84e38df Make it compile under -pedantic. 1999-08-03 10:18:27 +00:00
Andy Polyakov
4c22909e31 Extra i386+gcc bn_div.c tune-up featuring inline division and saving 
the remainder left in %edx. Here is the resulting performance improvement
matrix (improvement as a result of this *and* previous tune-up committed
two days ago). The results were obtained by profiling the "div" part of
the crypto/bn/bnspeed.c.

CPU	BN_div	bn_div_words	overall	comment
------------------------------------------------------------------------
PII	+16%	accumulated by	+2-3%	PII multiplies damn fast! Taking
		inlining		multiplication out of the loop
					didn't make too much difference.
					Eliminating of the multiplication
					involved in remainder calculation
					is the major factor.

Pentium	+45%	accumulated by	+7-9%	mull isn't that fast and replacing
		inlining		multiplications with additions in
					the loop has more visible effect:-)

MIPS	+75%	+12%		+20-25%	In addition to the taking mults
R10000					out of the loop (giving 12% in the
					asm/mips3.s) three mults were
					eliminated in BN_div.

Alpha	+30%	+50%		+10-15%	Same as above. But remember that
EV4					bn_div_words is a C implementation.
					It takes 4 Alpha mults in C to do
					the same thing as 1 MIPS mult in
					assembler does. So the effect (50%)
					is more impressive. But not the
					overall one... Well, if Alpha
					bn_mul_add would be implemented
					in assembler overall improvement
					would be closer to MIPS...
 1999-07-31 23:27:41 +00:00
Bodo Möller
8d85b33eb5 by request: let BN_dup(NULL) just return NULL 1999-07-30 19:22:57 +00:00
Andy Polyakov
0dd25e3606 Bignum division tune-up. Idea is to move multiplications in front of 
loop body and replace 'em with addition/subtraction.
 1999-07-30 11:43:43 +00:00
Andy Polyakov
5965902e6e when invoking bn_*_comba[48] result->top wasn't always set correctly. 1999-07-27 09:36:59 +00:00
Andy Polyakov
1656ef2997 SPARC Solaris config updates. 
./config sences whole range of SPARC instruction sets. Do note that
it favors Sun C now if both gcc and cc 4.2 or later are present!
 1999-07-25 22:25:12 +00:00
Andy Polyakov
a0618e3e5e Added support for SPARC Linux. 1999-07-25 15:13:49 +00:00
Andy Polyakov
fccbb9b34f - performance retunes, v8plus bn_*_comba routines are reimplemented; 
- support for GNU assembler (read SPARC Linux);
 1999-07-25 12:34:30 +00:00
Ulf Möller
40e29b1976 Remove obsolete files. 1999-07-22 21:57:41 +00:00
Andy Polyakov
cea538154b Get rid of redundant multiplications in bn_div_words. 1999-07-21 13:53:01 +00:00
Andy Polyakov
7363455fac MIPS III/IV assembler module is reimplemented. 1999-07-20 15:50:20 +00:00