Andy Polyakov [Sun, 5 Sep 1999 14:17:42 +0000 (14:17 +0000)]
SHA clean-up Intel assembler companion.
I've chosen to nest two functions in order to save about 4K. As a result
s1-win32.asm doesn't look right (nested PROC/ENDP SEGMENT/ENDS) and it's
probably impossible to compile. I assume I have to reconsider... But not
today...
Andy Polyakov [Sun, 5 Sep 1999 12:42:04 +0000 (12:42 +0000)]
SHA clean-up and (LP64) tune-up.
"Clean-up" stands for the fact that it's using common message digest
template ../md32_common.h and sha[1_]dgst.c are reduced down to
'#define SHA_[01]' and then '#include "sha_locl.h"'. It stands "(LP64)"
there because it's 64 bit platforms which benefit most from the tune-up.
The updated code exhibits 40% performance improvement on IRIX64
(sounds too good, huh? I probably should double check if it's not
some cache trashing that was holding it back before), 28% - on
Alpha Linux and 12% - Solaris 7/64.
Andy Polyakov [Sat, 28 Aug 1999 13:07:51 +0000 (13:07 +0000)]
RIPEMD160 shape-up Intel assembler companion. Cycle counter benchmarks
went down from 1050 to 921 cycles on Pentium II. I haven't checked the
figures on Pentium yet.
Bodo Möller [Wed, 18 Aug 1999 17:14:42 +0000 (17:14 +0000)]
Fix horrible (and hard to track down) bug in ssl23_get_client_hello:
In case of a restart, v[0] and v[1] were incorrectly initialised.
This was interpreted by ssl3_get_client_key_exchange as an RSA decryption
failure (don't ask me why) and caused it to create a _random_ master key
instead (even weirder), which obviously led to incorrect input to
ssl3_generate_master_secret and thus caused "block cipher pad is
wrong" error messages from ssl3_enc for the client's Finished message.
Arrgh.
Add functions to allow extensions to be added to certificate requests.
Modify obj_dat.pl to take its files from the command line. Usage is now
perl obj_dat.pl objects.h obj_dat.h
this should avoid redirection shell escape problems under Win32.
Bodo Möller [Tue, 10 Aug 1999 11:36:22 +0000 (11:36 +0000)]
Updates.
Prototypes and constant declarations for non-copying reads and writes for
BIO pairs (which is totally untested as of now, so I don't yet commit
the actual source code, but reserve the numbers to avoid conflicts).
Support for parsing of certificate extensions in PKCS#10 requests: these are
used by things like Xenroll. Also include documentation for extendedKeyUsage
extension.
Ok, I'll do a few more final platform tests and then I'll roll the 0.9.4
tarball baby for us. PLEASE DO NOT COMMIT ANYTHING UNTIL YOU SEE MY FINAL
COMMITS TO `STATUS' INDICATING THAT WE'VE REACHED 0.9.5-dev.
| revision 1.109
| date: 1999/06/24 19:39:23; author: ulf; state: Exp; lines: +0 -1
| Use the same CPU recogition method for FreeBSD 3 as for the other FreeBSD
| versions. (The FreeBSD and FreeBSD-elf Configure entries were identical.)
Because Ulf overlooked the ELF vs. a.out difference, i.e. the two entries were
_NOT_ identical! This way OpenSSL 0.9.4 at least compiles again under FreeBSD
2 and 3. For 0.9.5-dev I'll try to fix the stuff in a more general way to also
allow FreeBSD 4 and other variants.
I don't think this is a real showstopper. Our internal verify procedure lacks
even more and I consider this more a useful feature request than a release
showstopper.
Bodo Möller [Sat, 7 Aug 1999 02:51:10 +0000 (02:51 +0000)]
New option "-crlf" to s_client and s_server which tells them to convert
LFs into CRLFs when forwarding data from stdin to the TLS connection.
This is necessary for properly talking HTTP.
Because of the code freeze this change is by default disabled for now;
without -DAPPS_CRLF, the code is exactly as before.
Bodo Möller [Fri, 6 Aug 1999 11:18:44 +0000 (11:18 +0000)]
Dont' assume that something starting with '-' is a filename --
"openssl gendsa -help" now prints the usage summary, not error
messages that now file -help was found.
Configure fix for Windows: under Windows+perl "system" ends up going via the
command shell which means that redirecttion needs backslashes in the paths.
Andy Polyakov [Sat, 31 Jul 1999 23:27:41 +0000 (23:27 +0000)]
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...