GNOME Bugzilla – Bug 386909
[ffdec_h264] crash watching diggnation episode
Last modified: 2011-06-28 23:33:59 UTC
Version: 2.16.2 What were you doing when the application crashed? I was watching an episode of Diggnation (large H.264) when it suddenly crashed. Distribution: Ubuntu 6.10 (edgy) Gnome Release: 2.16.1 2006-10-02 (Ubuntu) BugBuddy Version: 2.16.0 Memory status: size: 205754368 vsize: 0 resident: 205754368 share: 0 rss: 70447104 rss_rlim: 0 CPU usage: start_time: 1166383953 rtime: 0 utime: 490 stime: 0 cutime:466 cstime: 0 timeout: 24 it_real_value: 0 frequency: 8 Backtrace was generated from '/usr/bin/totem' (no debugging symbols found) Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1". (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1226127696 (LWP 12661)] [New Thread -1363711072 (LWP 12674)] [New Thread -1349764192 (LWP 12673)] [New Thread -1341277280 (LWP 12672)] [New Thread -1332479072 (LWP 12671)] [New Thread -1310266464 (LWP 12669)] [New Thread -1292248160 (LWP 12668)] [New Thread -1277600864 (LWP 12665)] [New Thread -1269130336 (LWP 12664)] (no debugging symbols found) 0xffffe410 in __kernel_vsyscall ()
+ Trace 94930
Thread 6 (Thread -1310266464 (LWP 12669))
gstreamer issue, reassigning Thanks for taking the time to report this bug. Unfortunately, that stack trace is missing some elements that will help a lot to solve the problem, so it will be hard for the developers to fix that crash. Can you get us a stack trace with debugging symbols? Please see http://live.gnome.org/GettingTraces for more information on how to do so. Thanks in advance!
OK, is this any better? Distribution: Ubuntu 6.10 (edgy) Gnome Release: 2.16.1 2006-10-02 (Ubuntu) BugBuddy Version: 2.16.0 Memory status: size: 208658432 vsize: 0 resident: 208658432 share: 0 rss: 72859648 rss_rlim: 0 CPU usage: start_time: 1166397046 rtime: 0 utime: 528 stime: 0 cutime:500 cstime: 0 timeout: 28 it_real_value: 0 frequency: 12 Backtrace was generated from '/usr/bin/totem' (no debugging symbols found) Using host libthread_db library "/lib/tls/i686/cmov/libthread_db.so.1". (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread -1226021200 (LWP 18974)] [New Thread -1359135840 (LWP 18987)] [New Thread -1350743136 (LWP 18986)] [New Thread -1342256224 (LWP 18985)] [New Thread -1333392480 (LWP 18984)] [New Thread -1311179872 (LWP 18982)] [New Thread -1293161568 (LWP 18981)] [New Thread -1277494368 (LWP 18978)] [New Thread -1269023840 (LWP 18977)] 0xffffe410 in __kernel_vsyscall ()
+ Trace 94988
Thread 6 (Thread -1311179872 (LWP 18982))
The problem appears to be buried somewhere in the middle of the ffmpeg module. If you were running feisty, this would be a bit easier. As it is, I *can* give you a procedure that should work, but I would advise against doing this unless you really know what you're doing with apt-get. Effectively what this will try and do is only upgrade the gstreamer ffmpeg module to the feisty version and see if you can reproduce this still. If you're sure, do the following. Add deb http://archive.ubuntu.com/ubuntu/ feisty main restricted universe deb http://people.ubuntu.com/~pitti/ddebs feisty main universe to your /etc/apt/sources.list and then install gstreamer0.10-ffmpeg-dbgsym. This *may* try and pull in various other dependancies, and you're going to have to use your best judgement here to avoid having your entire setup upgraded to feisty. You might now not be able to reproduce the bug, in which case we can close this as fixed in the later version. Otherwise, you should now be able to get a better stacktrace. Oh, and after you've done all of that, I would advise removing the two added lines from your /etc/apt/sources.list again.
All right, I bit the bullet and just upgraded to Feisty completely (which has a few improvements that was worth the upgrade anyway). I can still get Totem to crash, but Bug Buddy doesn't come up when I do this now. To reproduce, I just open up episode 76 of diggnation (the large H.264) and seek through towards the end quickly. The video window turns grey for a brief moment, then it crashes. Anyhow, is there another way to get that backtrace?
(In reply to comment #4) > Anyhow, is there another way to get that backtrace? Yes. See http://live.gnome.org/GettingTraces/Details#gdb-not-yet-running
Created attachment 78721 [details] Ubuntu Crash Report
I was able to crash it again, but this time the Ubuntu crash program opened up. I've attached the smaller file it produced here (the large one was 13 MB, too large to upload).
*** Bug 388120 has been marked as a duplicate of this bug. ***
*** Bug 390069 has been marked as a duplicate of this bug. ***
*** Bug 395418 has been marked as a duplicate of this bug. ***
This is weird, the first backtraces shows... that it's using h261 ? (does that even exist ?). Justin, could you provide a debug log that would help us better figure out what's wrong ? # GST_DEBUG=3,ffmpeg:5 totem yourfile.name > log 2>&1 [wait for it to crash] # bzip2 log And then attach log.bz2 to this bug report.
Created attachment 93654 [details] Debug Log
I hope that helps. The Bziped one was over 1 MB, so I couldn't attach it. I used 7-zip instead, which compressed it better.
it's not syncing correctly after a seek, making ffmpeg crash.
Is this issue still unresolved? We're seeing it appear with the latest source using the following pipeline: videotestsrc ! ffmpegcolorspace ! x264enc bitrate=300 ! rtph264pay ! rtph264depay ! ffdec_h264 ! ffmpegcolorspace ! autovideosink
Forgot to mention that we're running this on Windows XP.
That pipeline works fine for me with latest GIT. Does it work for you too now? If not, could you get a new backtrace?
Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for. Thanks!