This bug was filed from the Socorro interface and is report bp-6ab9b81c-6f1a-4891-be44-a9dfd2110216 . ============================================================= 0 xul.dll nsBuiltinDecoderStateMachine::Run content/media/nsBuiltinDecoderStateMachine.cpp:991 1 xul.dll nsThread::ProcessNextEvent xpcom/threads/nsThread.cpp:633 2 xul.dll nsThreadStartupEvent::Run xpcom/threads/nsThread.cpp:207 3 nspr4.dll _PR_NativeRunThread nsprpub/pr/src/threads/combined/pruthr.c:426 4 nspr4.dll pr_root nsprpub/pr/src/md/windows/w95thred.c:122 5 mozcrt19.dll _callthreadstartex obj-firefox/memory/jemalloc/crtsrc/threadex.c:348 6 mozcrt19.dll __dllonexit obj-firefox/memory/jemalloc/crtsrc/onexit.c:276 7 mozcrt19.dll _threadstartex obj-firefox/memory/jemalloc/crtsrc/threadex.c:326 8 ntdll.dll ntdll.dll@0x39f01 I've hit this twice now in today's build when browsing random sites -- not sure steps to reproduce yet.

+ nsCOMPtr<nsIRunnable> event = new ShutdownThreadEvent(mDecoder->mStateMachineThread); + NS_DispatchToMainThread(event, NS_DISPATCH_NORMAL); + mDecoder->mStateMachineThread = nsnull; This part of the patch in bug 631058 interacts badly with nsBuiltinDecoder::Stop(): if (mStateMachineThread) mStateMachineThread->Shutdown(); This is the mechanism used to ensure that the state machine thread is no longer running before the decoder is destroyed. However, by setting mStateMachineThread to nsnull before the ShutdownThreadEvent executes, the check in Stop() will get skipped, even though the state machine thread is still potentially active.

It is #4 top crasher in today's build.

Backed out.

Would have been a hardblocker, but fixed anyhow.

There have been no recent crash reports with Firefox 4.0b12pre > 20110216.