You're correct, it seems to be a compiler optimization issue. Forcing the
openssl build to use -O1 instead of -O3 (-O2 also fails) resolves the issue.
This does not appear to be a problem with the same version of the compiler
on another 64-bit platform, Linux x86_64, so at first blush it appears that
gcc (4.x only?) has issues with 64-bit ABIs on IA64 (the same problem
occurred on Linux IA64 as well).
I suppose the only thing to be done on the openssl side of things, if
anything, is to adjust the linux-ia64 and hpux64-ia64-gcc targets to use -O1
instead of -O3.
If we were doing this, then all config lines would have minimum
optimization levels by now. For the record it should be noted that there
is possibility for a bug in source code that is exposed by newer
compiler, but then such bug is commonly observed on several *different
hardware* platforms. E.g. similar failure on ia64 and parisc2 would
deserve closer look than "tell them to drop optimization level and move on."
To what extent is openssl tested on different platforms and with which
compilers for each release?
There is no fixed list. If particular platform is tested, then vendor
compiler is preferred, whichever available at the $PATH. A.
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [email protected]
Automated List Manager [EMAIL PROTECTED]
