Summary: | ThreadSanitizer: Data race in WebCore::ScrollingThread::initializeRunLoop | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | David Kilzer (:ddkilzer) <ddkilzer> | ||||
Component: | WebCore Misc. | Assignee: | David Kilzer (:ddkilzer) <ddkilzer> | ||||
Status: | CLOSED INVALID | ||||||
Severity: | Normal | CC: | ap, cmarcelo, commit-queue, jamesr, koivisto, luiz, simon.fraser, tonikitoo, webkit-bug-importer | ||||
Priority: | P2 | Keywords: | InRadar | ||||
Version: | WebKit Nightly Build | ||||||
Hardware: | Unspecified | ||||||
OS: | Unspecified | ||||||
See Also: | https://bugs.webkit.org/show_bug.cgi?id=157527 | ||||||
Attachments: |
|
Description
David Kilzer (:ddkilzer)
2016-05-08 04:40:37 PDT
Created attachment 278361 [details]
Patch v1
Comment on attachment 278361 [details]
Patch v1
Seems a little bit messy the way this function combines Mac-specific code, some of which is guarded by #if PLATFORM(COCOA) and some of which is not, with platform independent code.
Comment on attachment 278361 [details] Patch v1 Clearing flags on attachment: 278361 Committed r200554: <http://trac.webkit.org/changeset/200554> All reviewed patches have been landed. Closing bug. Rolled out r200554 in r200561. Committed r200561: <http://trac.webkit.org/changeset/200561> Turns out the thread leak is for a separate issue (that reproduces when using --run-singly with "run-webkit-tests -1"). This issue is about a data race in WebCore::ScrollingThread::initializeRunLoop. Yep, this was a false positive due to use of bmalloc. Adding --additional-env-var=“Malloc=1” to the run-webkit-tests invocation makes these false positives go away. blame has an explicit check for injected ASan library, we should add the same for TSan. |