On Mon, 2009-12-21 at 09:45 +0800, Raymond Yau wrote:
2009/12/21 Chris cpollock@embarqmail.com
On Sun, 2009-12-20 at 07:47 +0800, Raymond Yau wrote:
I can reproduce this bug on pulseaudio-0.9.14 by
aplay -v --period-size=32 any_rate_not_equal_44100Hz.wav
Anyway to stop this from happening? When it does happen it's causing my drive to thrash for about 4-5 minutes or longer as shown below:
Dec 20 14:35:50 localhost pulseaudio[4331]: ratelimit.c: 271 events suppressed Dec 20 14:35:50 localhost pulseaudio[4331]: asyncq.c: q overrun, queuing locally
Dec 20 14:44:15 localhost pulseaudio[4331]: asyncq.c: q overrun, queuing locally
you have to provide a test case which can reproduce this bug and a full pulseaudio log which recorded how this bug occur
http://www.pulseaudio.org/wiki/Community#BugsPatchesTranslations
If autospawning is enabled (which it now is by default) you might have to race against it when restarting PA, so it might be a good idea to issue "pulseaudio -k ; pulseaudio -vvvvv" in a single command
In my case , I suspect the latency and watermark calculation did not get the correct watermark/sleeping time when the alsa application using small period size (128 bytes is the alsa-pulse-plugin 's minimum period size ) since the error "asyncq.c: q overrun, queuing locally" does not occur when using the normal period_size of those applications (e.g. aplay , mplayer work quite well ) or rewind too much since snd_pcm_rewind() is a new function which is only used by PA server
I have the output of the above command, where do I post it?