[vlc-devel] possible race condition
xxcv07 at gmail.com
Sat May 16 03:26:06 CEST 2009
Someone broke my trac account and I can't post a ticket although I can
I recently encountered this problem where goom plugin will crash if gcc
optimized vlc binary.
[0x5cdb59c] mpgatofixed32 audio output debug: libmad error: bad
[0x5cdbb84] main generic debug: thread started
[New thread 4224.0x1834]
[New thread 4224.0x15e4]
[New thread 4224.0x1934]
[New thread 4224.0x1134]
Program received signal SIGSEGV, Segmentation fault.
[Switching to thread 4224.0x1134]
0x669984d7 in zoom_filter_xmmx ()
#0 0x669984d7 in zoom_filter_xmmx ()
#1 0x66982ed7 in goom_update ()
#2 0x66981d38 in Thread (p_this=0x5cdbb84) at goom.c:359
#3 0x6a5c1c81 in thread_entry (data=0x5ce4900) at misc/threads.c:1061
#4 0x6a5c2447 in vlc_entry (data=0x5e51838) at misc/threads.c:741
It appears that the problem exist in the audio filter part of vlc. It
sometimes will crash sometimes not,
but most of the time will crash when goom visualization is called. As
you can see it segfaults shortly after
calling goom_update(). Seems likely that there's a problem with the
libvlccore I think.
I also encountered this similar problem when scaletempo gets a p_filter
NULL pointer at different line of the
scaletempo module. p_filter becomes NULL before the module could
It is strange that the library thread got called too late. Maybe who is
responsible of the threading part of the code
should polish it up a bit and stop been a slacker! Same to the person
who is responsible for gui part of the code (go to hell).
Sorry for saying this but I get cyber (bullied) abused on trac tickets
and that's enough. If you don't think it is a bug or you'd
want to ignore tickets and release "shit" then whatever, like I ever
cared about your project? Been wasting too much of
my time helping you fixing (most stupid mistakes you made) bugs that
you'd never fix.
I don't know what you're thinking but I'm just a "user" wants to report
bugs, not a programmer.
More information about the vlc-devel