diff options
author | 2016-02-21 13:36:50 -0800 | |
---|---|---|
committer | 2016-02-21 13:36:50 -0800 | |
commit | 03762fea75b0fe34ac0375407b83b6ae52e453b4 (patch) | |
tree | 81916dad8832c11247a1ed0686bfb210a9c113bd /samplecode/SampleWritePixels.cpp | |
parent | 7c249e531900929c2fe2cdde76619fa6d2538c49 (diff) |
fix race: this kCacheSize is probably not meant to be static
TSAN sees us racing to intialize the static const kCacheSize.
This static const value depends on the runtime value image->getSize(), so there
is a race to set it. I think this is unintentionally copy-paste from the other
tests that use a constant kCacheSize. It's weird to intialize a constant based
on the first call to test_internal_purge(), though luckily we're always calling
it with same sized images today.
See these TSAN failures:
https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-TSAN/builds/4937
https://uberchromegw.corp.google.com/i/client.skia/builders/Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-TSAN/builds/4940/steps/dm/logs/stdio
TBR=robertphillips@google.com
BUG=skia:
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1720743002
Review URL: https://codereview.chromium.org/1720743002
Diffstat (limited to 'samplecode/SampleWritePixels.cpp')
0 files changed, 0 insertions, 0 deletions