diff options
author | 2014-06-02 13:23:49 -0700 | |
---|---|---|
committer | 2014-06-02 13:23:49 -0700 | |
commit | be7de4fc4229b70b0954f3ea5e3b053dae4c1a9d (patch) | |
tree | ed79e35bed27d88478cb4dcd750dff0761bdf6ad /tools | |
parent | c77392ed58ec78ab19fa0e3ff99fb8110854fba2 (diff) |
Update TSAN suppressions.
TSAN suppressions need to be on their own line, even free of comments.
Also, temporarily add SK_ANNOTATE_UNPROTECTED_READ to sk_acquire_load.
Will remove this when we're done iterating on SkBarriers_tsan.h: TSAN
has an atomic load that makes the annotation moot.
BUG=skia:
R=bungeman@google.com, mtklein@google.com
Author: mtklein@chromium.org
Review URL: https://codereview.chromium.org/308073013
Diffstat (limited to 'tools')
-rw-r--r-- | tools/tsan.supp | 18 |
1 files changed, 13 insertions, 5 deletions
diff --git a/tools/tsan.supp b/tools/tsan.supp index d87f48609f..aab8211cec 100644 --- a/tools/tsan.supp +++ b/tools/tsan.supp @@ -1,12 +1,20 @@ # Suppressions for Thread Sanitizer +# +# CAREFUL! Comments must go on their own line or your suppressions will silently fail. # WebP races (harmlessly) choosing function pointers for SIMD versions of some of its functions. race:third_party/externals/libwebp # skia:2459 Seemingly misdiagnosed use-after-free, having something to do with software GL drivers. -race:SkGLContextHelper::init # We don't always get swrast_dri.so in the stack or we'd use that. +# We don't always get swrast_dri.so in the stack or we'd use that. +race:SkGLContextHelper::init -# For now assume anything using SkOnce and calls on SkRefCntBase (ref, unref, unique) are safe. -# TODO(mtklein): teach TSAN about SkOnce, SkRefCnt and SkSpinlock correctly. -race:SkOnce -race:SkRefCntBase +# Threadsafe, should be ported to SkLazyPtr. +race:SkFontHost_FreeType + +# Not threadsafe, should be fixed. +race:SkFontHost_fontconfig_ref_global +race:SkString::RefRec + +# Not threadsafe, should be deleted. +race:SkPDF |