Vorbis dev - Jan 2002 - ogg123 1.0rc3 thread issue

As I have mentioned here before, ogg123 from 1.0rc3 segfaults after
playing a file on FreeBSD and OpenBSD.  This happens with both the
native pthread and with GNU pth.  I have finally managed to isolate
the problem.

ogg123 crashes inside pthread_cond_wait(), specifically in the
COND_WAIT() call in buffer_wait_for_empty().

*If* I understand the control flow correctly, the decoder thread
waits on a condition variable here that has already been removed
by the other thread winding down in buffer_thread_cleanup().  This
would result in "undefined behavior".

Removing the pthread_*_destroy() calls from buffer_thread_cleanup()
eliminates the crash.

--- ogg123/buffer.c.orig	Thu Dec 20 01:24:53 2001
+++ ogg123/buffer.c	Wed Jan 16 03:26:15 2002
@@ -88,9 +88,11 @@
 
   /* Cleanup thread data structures */
   pthread_mutex_unlock(&buf->mutex);
+#if 0	/* KLUDGE: disabling this keeps ogg123 from crashing */
   pthread_mutex_destroy(&buf->mutex);
   pthread_cond_destroy(&buf->playback_cond);
   pthread_cond_destroy(&buf->write_cond);
+#endif
 }

<p>
-- 
Christian "naddy" Weisgerber                         
naddy@mips.inka.de

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

Christian Weisgerber wrote:
> 
> As I have mentioned here before, ogg123 from 1.0rc3 segfaults after
> playing a file on FreeBSD and OpenBSD.  This happens with both the
> native pthread and with GNU pth.  I have finally managed to isolate
> the problem.
Hi Christian, list,

I checked a fix for this problem into xiph.org cvs.  Only tested on OpenBSD/x86
(and regression tested on Linux/ppc), so you might want to test it on FreeBSD.

On OpenBSD 3.0, we still get bitten by the bug in libc_r where signals arriving
while a thread that doesn't handle them runs, don't get delivered to the
thread
that _does_ handle them, either.  In short, update libc_r from cvs and
be happy.
Otherwise, you won't be able to pause (^Z), skip to next song (^C), or quit
(^C^C).  You can always stop ogg123 with a SIGQUIT (^\), of course.

Btw, why do the BSD ports remove -O from the Makefiles?  It's actually
needed
(for libvorbis, at least).

Have fun,

<p>Segher

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to
'vorbis-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.