Bug 136073 – [GStreamer] crash after attempt to unlock mutex that was not locked

• Component: Media
• Status: RESOLVED WORKSFORME

[Michael Catanzaro]

Loading http://www.vox.com/2014/8/19/6043409/a-local-public-defender-on-the-deeply-dysfunctional-ferguson-court always causes my web process to crash, after printing this warning:

Attempt to unlock mutex that was not locked

Here's a stack trace from a release build; I can post one from a debug build if you think that'd be helpful.

#0  0x00007fa21c907c39 in __GI_raise (sig=sig@entry=6)
    at ../nptl/sysdeps/unix/sysv/linux/raise.c:56
        resultvar = 0
        pid = 5602
        selftid = 5647
#1  0x00007fa21c909348 in __GI_abort () at abort.c:89
        save_stage = 2
        act = {__sigaction_handler = {sa_handler = 0x57288, 
            sa_sigaction = 0x57288}, sa_mask = {__val = {140331614257200, 
              140331614268192, 140334287738662, 140333761429509, 0, 44, 
              140334240468288, 140331614257200, 140331614268192, 724000, 
              140334287765893, 44, 140334241384445, 140334244350464, 44, 0}}, 
          sa_flags = 482920960, sa_restorer = 0x7fa21b5acdf0}
        sigs = {__val = {32, 0 <repeats 15 times>}}
#2  0x00007fa21b528735 in g_mutex_unlock_slowpath (mutex=<optimized out>, 
    prev=<optimized out>) at gthread-posix.c:1321
No locals.
#3  0x00007fa21b5283ee in g_mutex_unlock (mutex=<optimized out>)
    at gthread-posix.c:1344
        prev = <optimized out>
#4  0x00007fa170b158bf in gst_hls_demux_change_playlist (
    demux=demux@entry=0x7fa18004c030, max_bitrate=9343035)
    at gsthlsdemux.c:1673
        previous_variant = 0x7fa18004eb00
        current_variant = 0x7fa18004eb20
        old_bandwidth = 357000
        new_bandwidth = 724000
        __FUNCTION__ = "gst_hls_demux_change_playlist"
#5  0x00007fa170b198ad in gst_hls_demux_switch_playlist (demux=0x7fa18004c030)
    at gsthlsdemux.c:1755
        bitrate = 11678794
#6  gst_hls_demux_stream_loop (demux=0x7fa18004c030) at gsthlsdemux.c:1249
        end_of_playlist = 0
        err = 0x0
        __FUNCTION__ = "gst_hls_demux_stream_loop"
#7  0x00007fa21825b346 in gst_task_func (task=0x80e710) at gsttask.c:317
        lock = 0x7fa18004c210
        tself = 0x7fa180049230
        priv = 0x80e6c0
        __PRETTY_FUNCTION__ = "gst_task_func"
#8  0x00007fa21b50c0cc in g_thread_pool_thread_proxy (data=<optimized out>)
    at gthreadpool.c:307
        task = 0x7fa178001fa0
        pool = 0xc4e4b0
#9  0x00007fa21b50b745 in g_thread_proxy (data=0x7fa180049230) at gthread.c:764
        thread = 0x7fa180049230
#10 0x00007fa21ac84f33 in start_thread (arg=0x7fa1a5ffb700)
    at pthread_create.c:309
        __res = <optimized out>
        pd = 0x7fa1a5ffb700
        now = <optimized out>
        unwind_buf = {cancel_jmp_buf = {{jmp_buf = {140332251461376, 
                -2793351388160340356, 0, 0, 140332251462080, 140332251461376, 
                2774021971184963196, 2772018513804643964}, 
              mask_was_saved = 0}}, priv = {pad = {0x0, 0x0, 0x0, 0x0}, 
            data = {prev = 0x0, cleanup = 0x0, canceltype = 0}}}
        not_first_call = <optimized out>
        pagesize_m1 = <optimized out>
        sp = <optimized out>
        freesize = <optimized out>
#11 0x00007fa21c9c6ded in clone ()
    at ../sysdeps/unix/sysv/linux/x86_64/clone.S:111
No locals.

[Philippe Normand]

*** Bug 136074 has been marked as a duplicate of this bug. ***

[Brendan Long]

Hm I don't get this in an Arch Linux VM either :\

[Michael Catanzaro]

Arch would still be on glib 2.40.  The strict mutex behavior is new in glib 2.42 (to be released late next month).

[Philippe Normand]

This trace is very similar to the one in Bug 138606. Shall we dupe one of these bugs?

[Michael Catanzaro]

Yeah it looks the same, though that makes me wonder if I was right to suspect glib 2.42 since if Andreas is using WebKit jhbuild, surely glib is much older.

[Michael Catanzaro]

I do not see this anymore.