Created attachment 470247 [details] All-threads backtrace This is a sequel to bug #257551. With WebKitGTK 2.43.4 and GStreamer 1.22.10, the main thread deadlocks when calling WebCore::unregisterPipeline instead a call to WebCore::deinitializeGStreamer, causing the watchdog thread to crash the process after 10 seconds. However, the backtrace is different than it was in bug #257551. Backtrace attached. Notably, only the main thread is using GStreamer here, which seems unusual. In the many other GStreamer deadlocks I've reported, there are usually at least two GStreamer threads visible in the backtrace. Perhaps some other thread left the mutex locked by mistake? I've hit this crash three times in the past hour, but I don't know what website is triggering it.
And you have the fix from #257551 or not?
Yes, because that bug was closed on January 1 but I'm using WebKitGTK 2.43.4 which was released in February.
(In reply to Michael Catanzaro from comment #2) > Yes, because that bug was closed on January 1 but I'm using WebKitGTK 2.43.4 > which was released in February. I was referring to https://commits.webkit.org/275032@main which landed on February 20th.
*** This bug has been marked as a duplicate of bug 264824 ***