Michael Niedermayer <[email protected]> added the comment: On Tue, Jun 30, 2009 at 01:10:40AM +0000, Jason Garrett-Glaser wrote: > Another Facebook bug report; delete the problematic file once the bug is > fixed. > > File is "facebook_h264_crash_3.mp4". On both Windows and Linux, this file > doesn't directly cause a segfault, rather it corrupts the heap in a way that > results in a libc crash after decoding is finished. > > GDB: > > Starting program: ffmpeg_g -i facebook_h264_crash_3.mp4 -f null -an /dev/null > FFmpeg version SVN-r19217, Copyright (c) 2000-2009 Fabrice Bellard, et al. > configuration: --enable-libfaad --disable-demuxer=v4l --disable-demuxer=v4l2 > --disable-ffserver --disable-ffplay --enable-libxvid --enable-libamr-wb > --enable-libamr-nb --enable-libx264 --enable-pthreads --disable-shared > --enable-nonfree --enable-static --enable-libfaac --enable-gpl > --enable-libmp3lame --enable-libtheora --enable-libvorbis > --enable-libschroedinger --enable-libspeex --disable-stripping > libavutil 50. 3. 0 / 50. 3. 0 > libavcodec 52.31. 2 / 52.31. 2 > libavformat 52.34. 0 / 52.34. 0 > libavdevice 52. 2. 0 / 52. 2. 0 > libswscale 0. 7. 1 / 0. 7. 1 > built on Jun 19 2009 13:24:02, gcc: 4.0.1 20050727 (Red Hat 4.0.1-5) > > Seems stream 1 codec frame rate differs from container frame rate: 60000.00 > (60000/1) -> 29.97 (30000/1001) > Input #0, mov,mp4,m4a,3gp,3g2,mj2, from > '/home/jgarrett/local/www/scripts/motion/1090784506275.mp4': > Duration: 00:02:24.03, start: 0.000000, bitrate: 533 kb/s > Stream #0.0(und): Audio: aac, 44100 Hz, 2 channels, s16 > Stream #0.1(und): Video: h264, yuv420p, 320x240 [PAR 1:1 DAR 4:3], 29.97 > tbr, 30k tbn, 60k tbc > Output #0, null, to '/dev/null': > Stream #0.0(und): Video: rawvideo, yuv420p, 320x240 [PAR 1:1 DAR 4:3], > q=2-31, 200 kb/s, 90k tbn, 29.97 tbc > Stream mapping: > Stream #0.1 -> #0.0 > Press [q] to stop encoding > > <enormous cascade of errors omitted> > > Program received signal SIGSEGV, Segmentation fault. > [Switching to Thread 0x2aaaac18a5e0 (LWP 4916)] > 0x00002aaaabc18b46 in _int_free () from /lib64/libc.so.6 > > I'd post the Valgrind log, but it is quite seriously too long (thousands of > errors, tens of thousands of lines).
well, posting some selected parts of it (a subset of invalid writes) might be usefull > > ---------- > assignedto: michaelni I would prefer if people did not randomly assign things to me this is especially true for possibly security relevant things without confirming that i will work on the issue. The reason is that a set assignedto might prevent people who actually have time to work on it from doing it. yes all the h264 issues should be fixed but assigning them to me wont speed that up, actually ive decided not to look into any of them anytime soon but rather work on more long standing issues first. That makes the assignedto rather misleading ... either way, until i get past "Sending libavcodec/mpegaudio_parser.c Transmitting file data .svn: Commit failed (details follow): svn: Can't read from connection: Connection timed out" " i cant work on anything, i wonder what is wrong with svn ATM ... [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Dictatorship naturally arises out of democracy, and the most aggravated form of tyranny and slavery out of the most extreme liberty. -- Plato _____________________________________________________ FFmpeg issue tracker <[email protected]> <https://roundup.ffmpeg.org/roundup/ffmpeg/issue1238> _____________________________________________________
