Strange race condition with recent drivers and mplayer!

[pffft!]

For quite some revisions now,

mplayer trallala.wmv

crashes with:

MPlayer SVN-r32735-4.1.2 (C) 2000-2010 MPlayer Team

Playing /joint/Videos/macbookpro.asf.
ASF file format detected.
[asfheader] Audio stream found, -aid 1
[asfheader] Video stream found, -vid 2
VIDEO:  [WMV3]  320x240  24bpp  1000.000 fps  214.0 kbps (26.1 kbyte/s)
Clip info:
 title:  
 author:  
 copyright:  
 comments:  
==========================================================================
Opening video decoder: [dmo] DMO video codecs
DMO dll supports VO Optimizations 0 1
DMO dll might use previous sample when requested


MPlayer interrupted by signal 11 in module: init_video_codec
- MPlayer crashed by bad usage of CPU/FPU/RAM.
  Recompile MPlayer with --enable-debug and make a 'gdb' backtrace and
  disassembly. Details in DOCS/HTML/en/bugreports_what.html#bugreports_crash.
- MPlayer crashed. This shouldn't happen.
  It can be a bug in the MPlayer code _or_ in your drivers _or_ in your
  gcc version. If you think it's MPlayer's fault, please read
  DOCS/HTML/en/bugreports.html and follow the instructions there. We can't and
  won't help unless you provide this information when reporting a possible bug.

This used to work perfectly well, and still does so on some other _identical_ machines on which the nvidia linux drivers are *not* installed.

If I compile with '--enable-debug', I no longer get a crash. But it does not seem to be a compiler bug! If I run the optimized executable under gdb, it works again. Stracing mplayer also does the trick!

With ltrace I get:

pthread_mutex_lock(0x8b32d08, 8, 0x9b8b208, 0x9b8b368, 0x9b8b330) = 0
pthread_mutex_unlock(0x8b32d08, 8, 0x9b8b208, 0x9b8b368, 0x9b8b330) = 0
free(0x9b43d20)
... and a lot more of this ...
pthread_mutex_lock(0x9b44328, 0x9b443c0, 0, 0xffe103b8, 0x82cbf01) = 0
pthread_cond_signal(0x9b44368, 0x9b443c0, 0, 0xffe103b8, 0x82cbf01) = 0
pthread_mutex_unlock(0x9b44328, 0x9b443c0, 0, 0xffe103b8, 0x82cbf01 <unfinished ...>
+++ killed by SIGTRAP +++

I cannot upload any of my test files for copyright reasons, sorry. But they used to work before and are still playable on identical machined without the nvidia binary driver.

Any idea?

[cehoyos]

Originally Posted by pffft!

This used to work perfectly well, and still does so on some other _identical_ machines on which the nvidia linux drivers are *not* installed.

Could you post output from those machines? Or try mplayer -vc ffwmv3?

Carl Eugen

[pffft!]

Could you post output from those machines?

I thought the nvidia-bug-report.log will contain any detail you want to know. Anyway, I'm prepared to tell you everything else ...

On one of my working machines, it looks like this:

MPlayer SVN-r32735-4.1.2 (C) 2000-2010 MPlayer Team

Playing /media/disk/macbookpro.asf.
ASF file format detected.
[asfheader] Audio stream found, -aid 1
[asfheader] Video stream found, -vid 2
VIDEO:  [WMV3]  320x240  24bpp  1000.000 fps  214.0 kbps (26.1 kbyte/s)
Clip info:
 title:  
 author:  
 copyright:  
 comments:  
==========================================================================
Opening video decoder: [dmo] DMO video codecs
DMO dll supports VO Optimizations 0 1
DMO dll might use previous sample when requested
Decoder supports the following formats: YV12 YUY2 UYVY YVYU RGB8 RGB555 RGB565 R
GB24 RGB32 
Decoder is capable of YUV output (flags 0x1b)
Movie-Aspect is undefined - no prescaling applied.
VO: [xv] 320x240 => 320x240 Planar YV12 
Selected video codec: [wmv9dmo] vfm: dmo (Windows Media Video 9 DMO)
==========================================================================
==========================================================================
Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders
AUDIO: 44100 Hz, 2 ch, s16le, 48.0 kbit/3.40% (ratio: 6003->176400)
Selected audio codec: [ffwmav2] afm: ffmpeg (DivX audio v2 (FFmpeg))
==========================================================================
AO: [oss] 44100Hz 2ch s16le (2 bytes per sample)
Starting playback...
Selected font is fixed-width.
Selected font is fixed-width.
A:  49.8 V:  49.8 A-V:  0.000 ct: -0.046 1313/1313  4%  0%  0.3% 0 0

mplayer -vc ffwmv3

This works, but that was not quite the point, of course.

Addendum: Segfaulting is independent of the output device selected.

[pffft!]

I'm still not sure who is to blame. Maybe this is an mplayer issue? But then again, it works on other machines without the nvidia driver, which run the exact same software. And I made the nvidia ones even crash with this!

Here is what '-v -v -v' gives me: