It it possible that the main thread will modify the m_tap member by calling AudioSourceProviderAVFObjC::destroyMix(), while m_tap is being read on a secondary thread in AudioSourceProviderAVFObjC::process().
Created attachment 310493 [details] Patch
Comment on attachment 310493 [details] Patch Can we construct a test case?
(In reply to Darin Adler from comment #2) > Comment on attachment 310493 [details] > Patch > > Can we construct a test case? I have not been able to reproduce this. I believe the probability of two threads simultaneously accessing m_tap is (very) small. Thanks for reviewing!
Comment on attachment 310493 [details] Patch Clearing flags on attachment: 310493 Committed r217878: <http://trac.webkit.org/changeset/217878>
All reviewed patches have been landed. Closing bug.