Hide Forgot
+++ This bug was initially created as a clone of Bug #1104812 +++ pthread_cond_wait on i686 has unwind data which the libgcc unwinder (from the gcc package) does not handle correctly. Ongoing discussion is here: https://gcc.gnu.org/ml/gcc/2016-10/msg00165.html This issue is not present in rawhide because the new condvar implementation does not have this problem.
Upstream commit: https://gcc.gnu.org/viewcvs/gcc?view=revision&revision=241491
Hi Florian, could you please advise me how this bug could be tested? Thanks!
(In reply to Michael Petlan from comment #4) > Hi Florian, could you please advise me how this bug could be tested? Thanks! I think the only thing you can do is to look at the disassembly of libgcc (objdump -d --reloc) and see if there is a reference to _Unwind_GetIPInfo in the function __gcc_personality_v0.
This is from ./obj-x86_64-redhat-linux/x86_64-redhat-linux/32/libgcc/32/libgcc_s.so 000158f0 <__gcc_personality_v0>: 158f0: 55 push %ebp 158f1: b8 03 00 00 00 mov $0x3,%eax 158f6: 57 push %edi 158f7: 56 push %esi 158f8: 53 push %ebx ... [snip] ... 15a10: 89 7c 24 48 mov %edi,0x48(%esp) 15a14: 89 04 24 mov %eax,(%esp) -->15a17: e8 f4 c4 fe ff call 1f10 <_Unwind_GetIPInfo@plt> 15a1c: 83 7c 24 28 01 cmpl $0x1,0x28(%esp) 15a21: 89 c7 mov %eax,%edi 15a23: 83 df 00 sbb $0x0,%edi 15a26: 39 74 24 48 cmp %esi,0x48(%esp) The 64bit libgcc (./obj-x86_64-redhat-linux/x86_64-redhat-linux/libgcc/libgcc_s.so) looks similar: 0000000000011e20 <__gcc_personality_v0>: 11e20: 41 57 push %r15 11e22: b8 03 00 00 00 mov $0x3,%eax 11e27: 41 56 push %r14 11e29: 41 55 push %r13 11e2b: 41 54 push %r12 ... [snip] ... 11f45: 48 89 7c 24 50 mov %rdi,0x50(%rsp) 11f4a: 4c 89 f7 mov %r14,%rdi -->11f4d: e8 1e 0a ff ff callq 2970 <_Unwind_GetIPInfo@plt> 11f52: 83 7c 24 14 01 cmpl $0x1,0x14(%rsp) 11f57: 48 89 c3 mov %rax,%rbx 11f5a: 48 83 db 00 sbb $0x0,%rbx 11f5e: 4c 39 7c 24 50 cmp %r15,0x50(%rsp) (both from gcc-4.8.5-16.el7) When looking at the same thing in gcc-4.8.5-11.el7, I don't see the reference in the same function. --> VERIFIED.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2017:2094