[llvm-dev] Unstable XRay test on ARM

[Oleg Ranevskyy via llvm-dev]

Hi Renato, Dean, Serge,

Just looked into the code and wanted to share some thoughts.

This might be a compare_exchange_weak spurious failure. ARM is a weakly
ordered CPU, but I am not sure whether spurious failures are really
possible in a single threaded app. On the other hand, there is no other way
for FDRLogging_init to fail in such a way (return XRAY_LOG_UNINITIALIZED
instead of XRAY_LOG_INITIALIZED) without any extra output. This is also
true for the 2nd test failure in FDRLogging_finalize, which uses a weak
exchange too.

Probably, the weak exchange needs to be either replaced with a strong one
or looped with more detailed CurrentStatus checks.

Oleg

On Wed, Jan 25, 2017 at 2:02 PM, Renato Golin via llvm-dev <
llvm-dev at lists.llvm.org> wrote:

> Hi Dean/Serge,
>
> I just spotted this on our bots:
>
> First failure, unrelated commit:
> http://lab.llvm.org:8011/builders/clang-cmake-armv7-a15-full/builds/3190
>
> 'XRay-Unit :: unit/XRayFDRLoggingTest/FDRLoggingTest.Simple' FAILED
> llvm/projects/compiler-rt/lib/xray/tests/unit/fdr_logging_test.cc:55:
> Failure
>       Expected: FDRLogging_init(kBufferSize, kBufferMax, &Options,
> sizeof(FDRLoggingOptions))
>       Which is: 0
> To be equal to: XRayLogInitStatus::XRAY_LOG_INITIALIZED
>       Which is: 2
> [  FAILED  ] FDRLoggingTest.Simple (0 ms)
> ...
>  1 FAILED TEST
> ==11476==XRay instrumentation map missing. Not initializing XRay.
>
> Then a similar, but not identical error:
> http://lab.llvm.org:8011/builders/clang-cmake-armv7-a15-full/builds/3191
>
> 'XRay-Unit :: unit/XRayFDRLoggingTest/FDRLoggingTest.Simple' FAILED
> llvm/projects/compiler-rt/lib/xray/tests/unit/fdr_logging_test.cc:58:
> Failure
>       Expected: FDRLogging_finalize()
>       Which is: 2
> To be equal to: XRayLogInitStatus::XRAY_LOG_FINALIZED
>       Which is: 4
> [  FAILED  ] FDRLoggingTest.Simple (0 ms)
> ...
>  1 FAILED TEST
> ==11476==XRay instrumentation map missing. Not initializing XRay.
>
> Note 0->2 on the first, 2->4 on the second.
>
> And then, for no reason, it's green again:
> http://lab.llvm.org:8011/builders/clang-cmake-armv7-a15-full/builds/3192
>
> Looks like an unstable test to me. :)
>
> Can you guys have a look?
>
> Thanks!
> --renato
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170125/a38111fc/attachment.html>