I am running alsa-1.0.18-8.9 and pulseaudio-0.9.12-9.5, currently the newest versions from OpenSuSE 11.1. The architecture is AMD64.
When I suspend the machine to disk and have it wake up again, in about 1 of 2 cases I get a pulseaudio process that is apparently in a loop writing the same 2 error messages to syslog, at a rate of 1413 messages per minute. The process is a standard userspace process that can be killed easily - once you notice something is wrong, which can take time unless you have a loud hard disk.
Here is a brief excerpt of the syslog messages (easy to guess how it is going to continue…):
May 6 08:01:00 hostname pulseaudio: alsa-util.c: Got POLLERR from ALSA May 6 08:01:00 hostname pulseaudio: alsa-util.c: PCM state is RUNNING May 6 08:01:00 hostname pulseaudio: alsa-util.c: Got POLLERR from ALSA May 6 08:01:00 hostname pulseaudio: alsa-util.c: PCM state is RUNNING
Originally I posted this in the Pulseaudio Forum (Ticket URL: <#559 (Pulseaudio loops writing syslog messages after waking from suspend to disk) – PulseAudio](http://pulseaudio.org/ticket/559)) but was told it is an ALSA bug.
From the response in that forum:
This is an ALSA issue. Due to some reason ALSA signals us an error via POLLERR. We then try to reset the audio device via snd_pcm_drop() and snd_pcm_prepare() which seems to succeed but then results in POLLERR right again apparently.
My motherboard (with onboard audio, don’t know the chipset) is an Asus M3N78-EM AM2+.
I would be happy if this bug could be fixed.
Many thanks, Ned.