This is an archive of the discontinued LLVM Phabricator instance.

[fuzzer][test] Disable noasan-strncmp test for AArch64
ClosedPublic

Authored by antmo on Apr 14 2023, 5:42 AM.

Details

Summary

This test fails on several aarch64 bots

clang-aarch64-lld-2stage ( https://lab.llvm.org/buildbot/#/builders/185/builds/3525 )
clang-aarch64-full-2stage ( https://lab.llvm.org/buildbot/#/builders/179/builds/5904 )
clang-aarch64-sve-vla ( https://lab.llvm.org/buildbot/#/builders/197/builds/4519 )

Disable it while this is analyzed.

Diff Detail

Event Timeline

antmo created this revision.Apr 14 2023, 5:42 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 14 2023, 5:42 AM
antmo requested review of this revision.Apr 14 2023, 5:42 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 14 2023, 5:42 AM
Herald added a subscriber: Restricted Project. · View Herald Transcript

I dont get it for now. There is only one patch between the last success ( https://lab.llvm.org/buildbot/#/builders/197/builds/4518 ) and the first failure ( https://lab.llvm.org/buildbot/#/builders/179/builds/5904 ) : https://reviews.llvm.org/D147121 ( b356cd70a27d )
But this is very unlikely due to this patch, as it should have no effect on linux (and -pie is enabled when I reproduce the failure)

I suspect this test is sensitive to fluctuations in resources. With us running large shared machines that can happen a lot, despite fixing parts of it, like core count, for some containers. Potentially we could reproduce by running stress inside a container while running this test on its own.

compiler-rt/test/fuzzer/noasan-strncmp.test
1

You should use the newer syntax: https://github.com/llvm/llvm-project/commit/8a86860207f0d5c794a2b8c0b384001710e5d036

For example another test that excludes windows and aarch64:

UNSUPPORTED: darwin, target={{.*windows.*}}, target=aarch64{{.*}}

(I think darwin doesn't use it because the word darwin doesn't appear in the target triple)

Also you are right to use UNSUPPORTED not XFAIL, because when an XFAIL passes it becomes an XPASS, which is, itself, a failure.

If that makes any sense at all :)

DavidSpickett accepted this revision.Apr 14 2023, 8:29 AM

LGTM, do you have commit access yet?

This revision is now accepted and ready to land.Apr 14 2023, 8:29 AM
antmo added a comment.Apr 14 2023, 8:31 AM

LGTM, do you have commit access yet?

not yet ;)

This revision was landed with ongoing or failed builds.Apr 14 2023, 8:33 AM
This revision was automatically updated to reflect the committed changes.