[pulseaudio-tickets] [PulseAudio] #755: Pulseaudio randomly starts taking 100% cpu in mainloop polling

PulseAudio trac-noreply at tango.0pointer.de
Wed Jan 13 17:28:47 PST 2010


#755: Pulseaudio randomly starts taking 100% cpu in mainloop polling
------------------------+---------------------------------------------------
  Reporter:  lukehutch  |       Owner:  lennart
      Type:  defect     |      Status:  new    
 Milestone:             |   Component:  daemon 
Resolution:             |    Keywords:         
------------------------+---------------------------------------------------

Comment(by lennart):

 The gdb bt shows that PA is actually sleeping properly in the poll()
 loops. That's how it supposed to be.

 Hmm, can you trace back when that fd 29 was created that poll is woken up
 from? Just go backwards from a line like that and try to figure out where
 the fd 29 comes from.

 (this could some client simply keeping pa busy by creating connections in
 a loop)

-- 
Ticket URL: <http://pulseaudio.org/ticket/755#comment:2>
PulseAudio <http://pulseaudio.org/>
The PulseAudio Sound Server


More information about the pulseaudio-bugs mailing list