[Webkit-unassigned] [Bug 159396] New: REGRESSION(r202556): [GStreamer] Deadlock trying to play a video

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=159396

            Bug ID: 159396
           Summary: REGRESSION(r202556): [GStreamer] Deadlock trying to
                    play a video
    Classification: Unclassified
           Product: WebKit
           Version: WebKit Local Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Keywords: Gtk, Regression
          Severity: Normal
          Priority: P2
         Component: Platform
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: cgarcia at igalia.com
                CC: bugs-noreply at webkitgtk.org, pnormand at igalia.com,
                    slomo at coaxion.net

Created attachment 282687
  --> https://bugs.webkit.org/attachment.cgi?id=282687&action=review
Backtrace

It seems the gst_pad_query_convert introduced in r202556 is causing another deadlock in the gst locks mess. The convert function starts a new query on the pad that ends up in gst_video_decoder_sink_query_default() that tries to get the GST_VIDEO_DECODER_STREAM_LOCK. But that lock is taken by another thread that calls gst_video_decoder_finish_frame -> gst_video_decoder_clip_and_push_buf -> gst_pad_push and this ends up waiting in gst_queue_chain_buffer_or_list. This has nothing to do with HLS deadlocks, I just tried to play a youtube video embedded in another website. See the backtraces attached.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160704/b2808db6/attachment-0001.html>