[pulseaudio-discuss] Pulse audio echo module problem with Alsa

[Alexander E. Patrakov]

16.05.2014 12:45, Dheeraj Kumar Dodda wrote:
> alsa-source.c: Increasing minimal latency to 16.00 ms
> alsa-sink.c: Scheduling delay of 24.50 ms > 10.00 ms, you might want to investigate this to improve latency...
> alsa-sink.c: Increasing minimal latency to 26.00 ms
> alsa-source.c: Increasing minimal latency to 26.00 ms
> alsa-sink.c: Increasing wakeup watermark to 16.00 ms
> alsa-source.c: Increasing wakeup watermark to 16.00 ms
> alsa-sink.c: Increasing minimal latency to 36.00 ms
> alsa-source.c: Increasing minimal latency to 36.00 ms
> alsa-source.c: Increasing wakeup watermark to 26.00 ms
> alsa-sink.c: ALSA woke us up to write new data to the device, but there was actually nothing to write!
> alsa-sink.c: Most likely this is a bug in the ALSA driver '(null)'. Please report this issue to the ALSA developers.
> alsa-sink.c: We were woken up with POLLOUT set -- however a subsequent snd_pcm_avail() returned 0 or another value < min_avail.
> alsa-source.c: Increasing minimal latency to 46.00 ms
> alsa-source.c: Increasing wakeup watermark to 36.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 56.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 46.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 66.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 56.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 76.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 66.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 86.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 76.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 96.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 86.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 106.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing wakeup watermark to 96.00 ms
> I: [alsa-source-SGTL5000 SGTL5000-0] alsa-source.c: Increasing minimal latency to 116.00 ms
> alsa-source.c: Increasing wakeup watermark to 106.00 ms
> alsa-source.c: Increasing minimal latency to 126.00 ms
> alsa-source.c: Increasing wakeup watermark to 116.00 ms
> alsa-source.c: Increasing minimal latency to 128.00 ms
> alsa-source.c: Increasing wakeup watermark to 118.00 ms
> alsa-source.c: Overrun!

OK, so now we have a different error, thus confirming my hypothesis 
about insufficient CPU resources for echo cancellation. If your board 
supports that, try using Debian's "armhf" port on it, as it provides 
better performance for floating-point operations.

However, the above messages about latency are not normal. They are also 
not bugs in PulseAudio - it just complains about your kernel.

If you can, please run latencytop and see the cause of the "Scheduling 
delay" message there. Also please verify your rtkit setup.

As for "ALSA woke us up", it is clearly a kernel bug. Please fix your 
driver.

If you need to separate the two bugs (scheduling latency vs premature 
wakeup), try using a USB headset.

-- 
Alexander E. Patrakov