>>>>> I see a very strange bug in crypto/sparcv9cap.c. OpenSSL 1.0.0d checks >>>>> sparc capabilities with SIGILL signal. On sparc64 (both Linux and >>>>> solaris, with UltraSPARC III+ and T1 CPU's), SIGILL handler is called >>>>> and program terminates with SIGILL in _sparcv9_fmadd_probe: >>>>> >>>>> 00000001002a32d0<_sparcv9_fmadd_probe>: >>>>> 1002a32d0: 81 b0 0d 80 impdep1 108, %f0, %f0, %f0 >>>>> 1002a32d4: 85 b0 8d 82 impdep1 108, %f2, %f2, %f2 >>>>> 1002a32d8: 81 c3 e0 08 retl >>>>> 1002a32dc: 81 b8 04 40 impdep2 34, %f0, %f0, %f0<= here > > Question : if sparc processor cannot execute impdep2 instruction, > SIGILL is caught and program has to jump to signal handler.
... and the question is? I mean it sounds more like a statement rather than question... And it can be strengthened. If *a* processor cannot execute unimplemented instruction, modern OSes make it possible to arrange that execution flow is transferred to user-supplied code. BTW. As for question posed earlier. Why do I test for sparcv9 instructions in 32-bit mode? v9 instructions are perfectly usable even in 32-bit mode. And they *are* extensively used in assembler, even in 32-bit application context. Even compilers were observed to use them, also in 32-bit code. If your goal is to produce 32-bit binary that can be executed on wider range of SPARC-based systems, then you should be configuring for linux-sparcv8 (where the code in question is not even compiled). For reference. Why do I test for fmadd? Well, this is actually reserved for future, no code is using it now, so that this particular check can be omitted [for now]. >>>>> If I add printf() in signal handler, I see that it is called, and that >>>>> siglongjmp() works. With my printf(), my program doesn't abort with >>>>> SIGILL anymore but with SIGBUS (?!). >>>> Could you 'truss -v sigaction,sigprocmask apps/openssl version' and >>>> submit output? >>> I cannot. My server doesn't run Solaris but Linux/sparc. I just have >>> seen the same bug a long time ago on Solaris but I don't have any >>> solaris server anymore. >> Then run 'strace -v apps/openssl version'. > > Output is attached. But according to the output, the apps/openssl didn't crash, but caught *and* handled SIGILL. So it works at least in apps/openssl context. Do same against your application that crashes. Also do tell if there is anything special about your application, in particular what other libraries is it linked with. ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
