[Mplayer-users] Probable race condition issues preclude Mplayer from workingThu Apr 19 01:05:06 2001

[Arpi]

Hi,

> 	Hi.  I upgraded my PC recently with a new CPU/mobo to a 1Ghz Tbird.  But
> now mplayer goes into an infinite loop of somekind when it attempts to play
> video.  Tested the most recent mplayer CVS just to be sure it wasn't fixed
> recently.
> 
> -=- System info -=-
> Kernel version number: Linux shadowgate 2.2.19 #1 Sat Mar 31 19:48:50 CST
> 2001 i686 unknown
> Linux distribution and glibc version: Slackware 7.1 and glibc 2.1.3
> gcc and binutils version: gcc version 2.95.2 19991024 (release) and GNU
> assembler 2.9.1
> X11 version: XFree86 4.0.2
> Video card vendor/model: nVidia Riva TNT 1
> CPU vendor and type: AMD Thunderbird Athlon at 1,000Mhz
> -=- -=-
> 
> The infinite loop happens right after the final line in the attached
> play.log ('Start playing...').
Hmm. Sound card? Is it unchanged (both card and driver)?
Does it work with -nosound ?

>  This affects all media formats, so I think
> it's a race condition somewhere in the core.
There is no such code.. everything is done in a single process, no threads
no more processes. I'm sure that problem is caused by audio driver.
Try to play with audio parameters, like #undef HAVE_AUDIO_SELECT in config.h
(+recompile) and -noalsa.

> Perhaps proper thread interlocking or a more finely grained sleep routine
> would solve the issue?
There is an usleep() loop, maybe it's broken. Try with -noalsa, it disables
that code.

> Audio buffer size: 15872 bytes, delay: 0.496s
seems to be ok.


A'rpi / Astral & ESP-team

--
mailto:arpi at thot.banki.hu
http://esp-team.scene.hu

_______________________________________________
Mplayer-users mailing list
Mplayer-users at lists.sourceforge.net
http://lists.sourceforge.net/lists/listinfo/mplayer-users