Bug 113296 - [14 Regression] SPEC 2006 434.zeusmp segfaults on Aarch64 when built with -Ofast -march=native -flto
Summary: [14 Regression] SPEC 2006 434.zeusmp segfaults on Aarch64 when built with -Of...
Status: RESOLVED FIXED
Alias: None
Product: gcc
Classification: Unclassified
Component: target (show other bugs)
Version: 14.0
: P3 normal
Target Milestone: 14.0
Assignee: Not yet assigned to anyone
URL:
Keywords: needs-bisection, needs-reduction, wrong-code
Depends on:
Blocks: spec
  Show dependency tree / graph
 
Reported: 2024-01-09 17:48 UTC by Martin Jambor
Modified: 2024-01-12 12:09 UTC (History)
0 users

See Also:
Host: aarch64-linux
Target: aarch64-linux
Build:
Known to work:
Known to fail:
Last reconfirmed:

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Martin Jambor 2024-01-09 17:48:42 UTC 
Our Aarch64 benchmarker (armv8.2-a+crypto+fp16+rcpc+dotprod+ssbs)
signals that SPEC 2006 434.zeusmp segfaults at run-time when built
with -Ofast -march=native -flto and master revision
r14-7022-g34d339bbd0c1f5.


Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x40000002978f in ???
#1  0x403ff0 in ggen_
        at /home/gcc/buildworker/source/cpu2006/benchspec/CPU2006/434.zeusmp/build/build_peak_amd64-m64-mine.0000/ggen.f:762
#2  0x407a0f in setup_
        at /home/gcc/buildworker/source/cpu2006/benchspec/CPU2006/434.zeusmp/build/build_peak_amd64-m64-mine.0000/setup.f:1135
#3  0x40fc3b in mstart_
        at /home/gcc/buildworker/source/cpu2006/benchspec/CPU2006/434.zeusmp/build/build_peak_amd64-m64-mine.0000/mstart.f:301
#4  0x425ee3 in zeusmp
        at /home/gcc/buildworker/source/cpu2006/benchspec/CPU2006/434.zeusmp/build/build_peak_amd64-m64-mine.0000/zeusmp.fppized.f:620
#5  0x400d5f in main
        at /home/gcc/buildworker/source/cpu2006/benchspec/CPU2006/434.zeusmp/build/build_peak_amd64-m64-mine.0000/zeusmp.fppized.f:769


Unfortunately at the moment don't have another access to another
adequate Aarch64 machine to debug further and so I cannot provide more
information (and so the component "target" is likely bogus, sorry).
Comment 1 Martin Jambor 2024-01-12 12:09:20 UTC 
According to our buildbot results, this has resolved itself somewhen between 1 and 2 days ago.

I assume nobody wants to go an investigate what issue it was if it does not reappear, so let me close the bug.