I'm writtign this but I know you will probably have no idea how to fix it or what the problem is all I can say is they with a newer snapshot of openssl we have a problem with core dumping and it works when compiled with older versions.
Well the BN_CTX code has been changed recently and BN_POOL_get belongs to the new ctx code ...
Stack trace is as follows:
#0 0x08087522 in BN_POOL_get ()
(gdb) bt
#0 0x08087522 in BN_POOL_get ()
#1 0x0812ad18 in ?? ()
#2 0x0812b860 in ?? ()
#3 0x0812b5f0 in ?? ()
#4 0x08087419 in BN_CTX_get ()
#5 0x00000002 in ?? ()
#6 0x00000002 in ?? ()
#7 0x080d68f9 in ec_GFp_simple_point2oct ()
#8 0xb75ea19c in ?? ()
#9 0x00000030 in ?? ()
Previous frame inner to this frame (corrupt stack?)
(gdb) q
as the stack seems to be corrupted the stack trace isn't very reliable, but as the BN_CTX code has been rewritten it's not unlikely that the cause is somewhere in the ctx code
We are using the EC features.
that's good to hear ;-)
Guess it's possible our code is doing something wrong to screw up openssl but using a memory leak detection utility all leaks complaints are in openssl now. I haven't narrowed what call I make does this as of yet.
a code example that reproduce this bug would be interessting (btw: defining BN_CTX_DEBUG gives you a lot of perhaps useful information)
Nils
______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]