Hi Tom and all others,

what is that RSA fix you're talking about?
Recently I had some application crashes when creating lots of DSH
signatures on a 4 x UltraSparc-II (Solaris 8).

The crash always occurs in
#0  0xff0c1384 in _smalloc () from /usr/lib/libc.so.1
#1  0xff0c13cc in malloc () from /usr/lib/libc.so.1
#2  0xff0b5f6c in calloc () from /usr/lib/libc.so.1

while the stack above #2 always looks different so I suspect some
code in DSH to mess up the heap. Synchronizing the DSA_sign_setup()
and DSA_sign() calls alone did not help so I'd be interested in
what you fixed with RSA signatures.

Robert

P.S. Using RSA my application did not crash yet but that could
be a coincidence. Lately it doesn't crash with DSA either :-(


Subject:  DSA threading issue
From:     Tom Wu <[EMAIL PROTECTED]>
Date:     2002-08-30 19:39:30

The OpenSSL DSA implementation (as of 0.9.6g) appears to have the same
multithreading issue that the RSA implementation used to have, namely
that multiple threads simultaneously triggering the BN_MONT_CTX caching
results in a race condition.  My own testing suggests that the DSA code
can benefit from the same fix as was applied to RSA.

If anyone needs patches, I'll jump through the appropriate BXA hoops and
post them here.

Tom
--
Tom Wu
Principal Software Engineer
Arcot Systems
(408) 969-6124
"The Borg?  Sounds Swedish..."

______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [EMAIL PROTECTED]
Automated List Manager                           [EMAIL PROTECTED]

______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
Development Mailing List                       [EMAIL PROTECTED]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to