Re: [xine-devel] closing alsa driver freezes GUI

[Bastien N. <ha...@ha...>]

On Sun, 2004-03-21 at 13:25 +0100, Thibaut Mattern wrote:
> On Sat, 2004-03-20 at 21:59, Bastien Nocera wrote:
> > On Sat, 2004-03-20 at 02:09 +0000, Bastien Nocera wrote:
> > > On Fri, 2004-03-19 at 14:23, Bastien Nocera wrote:
> > > > On Fri, 2004-03-19 at 15:16, Mattern Thibaut wrote:
> > > > > Hi Daniel,
> > > > > 
> > > > > Citation de  Daniel Caujolle-Bert <seg...@cl...>:
> > > > > 
> > > > > > Hi Thibaut,
> > > > > > 
> > > > > > Le vendredi 19 Mars 2004 15:21, Mattern Thibaut a
> > > > > > =E9crit=A0:
> > > > > > > Citation de  Daniel Caujolle-Bert
> > > > > > <seg...@cl...>:
> > > > > > > > Hi J=FCrgen,
> > > > > > > >
> > > > > > > > 	can you try this patch ?
> > > > > > >
> > > > > > > why not a simple pthread_testcancel() at the end of the
> > > > > > loop ?
> > > > > > 
> > > > > > 	not needed, now we just not cancel the thread, but
> > > > > > finish it (with a mutex=
> > > > > >  in=20
> > > > > > a clean state ;-) ).
> > > > > 
> > > > > It looks really better now ! ;)
> > > > > That should fix Bastien's problem too.
> > > > 
> > > > I can test that tonight. I was getting absolutely nowhere with my
> > > > testing/debugging. Quite frustrating ;)
> > > > 
> > > > > Ready to release 1rc3ba ? ;)
> > > > 
> > > > I'll post some feedback tonight.
> > > 
> > > It works, it works, wheeee!
> > > Thanks Daniel (and the others who helped out on the bug)
> > > 
> > > My users will be happy :)
> > 
> > Well, it actually doesn't work properly :/
> 
> You don't describe the problem. We need more info.
> What's the problem now ?

Hang on exit:

(gdb) thread apply all bt
 
Thread 3 (Thread 64355248 (LWP 6923)):
#0  0x00a457a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
#1  0x00cb0550 in pthread_cond_timedwait@@GLIBC_2.3.2 ()
   from /lib/tls/libpthread.so.0
#2  0x0011c394 in metronom_sync_loop (this=0x83a4c40) at metronom.c:835
#3  0x00cad61c in start_thread () from /lib/tls/libpthread.so.0
#4  0x00b18c1a in clone () from /lib/tls/libc.so.6
 
Thread 2 (Thread 27921328 (LWP 6929)):
#0  0x00a457a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2
#1  0x00cb28ce in __lll_mutex_lock_wait () from /lib/tls/libpthread.so.0
#2  0x00caf963 in _L_mutex_lock_25 () from /lib/tls/libpthread.so.0
#3  0x01aa0a60 in ?? ()
#4  0x01aa0a18 in ?? ()
#5  0x0012a4d9 in xine_event_send (stream=0x88d95b8, event=0x83a52d4)
    at events.c:75
Previous frame identical to this frame (corrupt stack?)
#0  0x00a457a2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2

I also sometimes get crashes in xine_event_send() during the normal
course of use (ie. not only on exit).

> > I committed a patch that allows some front-ends to disable the event
> > thread. I think that the actual bug might be down to the alsa-lib
> > version, but as Totem and other front-ends relying on the software mixer
> > don't need it, saves a little bit of resources.
> > 
> > Daniel, let me know if it suits you. I attached the patch for the
> > record.
> > 
> > ---
> > Bastien Nocera <ha...@ha...> 
> > Remember that age and treachery will always triumph over youth and
> > ability.
---
Bastien Nocera <ha...@ha...> 
He was as lame as a duck. Not the metaphorical lame duck either, but a
real duck that was actually lame. Maybe from stepping on a land mine or
something.