New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 843589 link

Starred by 1 user

Issue metadata

Status: Duplicate
Merged: issue 831667
Owner:
Closed: May 2018
Cc:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 1
Type: Bug



Sign in to add a comment

LSAN failure: gin_unittests --gtest_filter=*ArgumentsTest.TestGetAll*

Project Member Reported by lukasza@chromium.org, May 16 2018

Issue description

This failure started pretty recently - in https://ci.chromium.org/buildbot/chromium.memory/Linux%20Chromium%20OS%20ASan%20LSan%20Tests%20%281%29/27475

I can repro some LSAN leak reports locally:
    $ cat out/lsan/args.gn 
    is_asan = true
    is_lsan = true
    is_debug = false  # Release build.
    enable_nacl = false
    use_goma = true

    $ ninja -C out/lsan -j 500 -l 25 gin_unittests
    ...

    $ ASAN_OPTIONS="detect_leaks=1 symbolize=1 external_symbolizer_path=$SRC/third_party/llvm-build/Release+Asserts/bin/llvm-symbolizer" out/lsan/gin_unittests --gtest_filter=*ArgumentsTest.TestGetAll*

But... the surprising issue that prevents me from bisecting is that I can also locally repro LSAN leaks going back for quite a while - e.g. at 9b7a2e09ec0532063b6cc14d4e4656898c6decb1 (green build https://ci.chromium.org/buildbot/chromium.memory/Linux%20Chromium%20OS%20ASan%20LSan%20Tests%20%281%29/27471) or 9b7a2e09ec0532063b6cc14d4e4656898c6decb1~2048
 
Labels: -Pri-3 Pri-1
Consistently red bots probably justifies treating this as Pri1
Owner: tzik@chromium.org
Status: Assigned (was: Untriaged)
tzik@ - I wonder if you could PTAL (as somebody who relatively recently touched gin/arguments_unittest.cc in a CL that seems related to base::BindRepeating which is mentioned in the failure log at https://logs.chromium.org/v/?s=chromium%2Fbb%2Fchromium.memory%2FLinux_Chromium_OS_ASan_LSan_Tests__1_%2F27475%2F%2B%2Frecipes%2Fsteps%2Fgin_unittests%2F0%2Flogs%2FArgumentsTest.TestGetAll%2F0).
Cc: h...@chromium.org sfiera@chromium.org
sfiera@ points out that this might be related to r558709 which might have enabled gin_unittests on the LSan bot (?).

Comment 4 by sfiera@chromium.org, May 16 2018

https://crrev.com/c/1061501 is already in-flight.
Mergedinto: 831667
Status: Duplicate (was: Assigned)
Great - thanks!

Sign in to add a comment