[pulseaudio-tickets] [PulseAudio] #865: Pulseaudio segfaults

[PulseAudio]

#865: Pulseaudio segfaults
-----------------------+----------------------------------------------------
  Reporter:  rabbit64  |       Owner:  lennart         
      Type:  defect    |      Status:  new             
 Milestone:            |   Component:  module-combine-*
Resolution:            |    Keywords:                  
-----------------------+----------------------------------------------------

Comment(by tanuk):

 So you've patched alsa-sink.c to define DEBUG_TIMING. Why? These
 "segfaults" are not segfaults, instead the daemon gets the debug trap
 signal.

 The check_left_to_play() function has this piece of code:

 {{{
 #ifdef DEBUG_TIMING
         PA_DEBUG_TRAP;
 #endif
 }}}

 Are you arguing that it's a bug to trigger the signal when DEBUG_TIMING is
 defined? I don't have an opinion myself, since it's not documented what
 DEBUG_TIMING means, so I don't know what's the intended behaviour - is
 triggering the signal a good thing to do or not.

 Anyway, this message is the reason why this code path is executed:

 {{{
 pulseaudio[14730]: (   3.543|   0.000) alsa-util.c: snd_pcm_avail()
 returned a value that is exceptionally large: 319392 bytes (1810 ms).
 pulseaudio[14730]: (   3.543|   0.000) alsa-util.c: Most likely this is a
 bug in the ALSA driver 'snd_emu10k1'. Please report this issue to the ALSA
 developers.
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c: snd_pcm_dump():
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c: Hardware PCM card 1
 'SB Live! 5.1 [SB0220]' device 0 subdevice 0
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c: Its setup is:
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   stream       :
 CAPTURE
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   access       :
 MMAP_INTERLEAVED
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   format       :
 S16_LE
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   subformat    : STD
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   channels     : 2
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   rate         : 44100
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   exact rate   : 44100
 (44100/1)
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   msbits       : 16
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   buffer_size  : 4096
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   period_size  : 2048
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   period_time  : 46439
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   tstamp_mode  :
 ENABLE
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   period_step  : 1
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   avail_min    : 2048
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   period_event : 1
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   start_threshold  :
 -1
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   stop_threshold   :
 4611686018427387904
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   silence_threshold: 0
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   silence_size : 0
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   boundary     :
 4611686018427387904
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   appl_ptr     : 49256
 pulseaudio[14730]: (   3.544|   0.000) alsa-util.c:   hw_ptr       :
 129104
 }}}

 So what to do? Should this ticket be closed? Should we remove the
 PA_DEBUG_TRAP line from check_left_to_play()?

-- 
Ticket URL: <http://pulseaudio.org/ticket/865#comment:3>
PulseAudio <http://pulseaudio.org/>
The PulseAudio Sound Server