Bug#726556: pulseaudio: mic is dead after suspend

[Stefan Krastanov]

Here are the results:

1.
`arecord` while pa is *not* suspended and while the mic is working
fine returns "device is busy"

```
 % arecord -vv -fdat -c 2 -D hw:0,0 test.wav
arecord: main:722: audio open error: Device or resource busy
```

2.
if the mic is working just before issuing the command
`pasuspender -- arecord -vv -fdat -c 2 -D hw:0,0 test.wav`
then `arecord` works and the sound from the mic is recorded

3.
if the mic is *not* working just before issuing the command
`pasuspender -- arecord -vv -fdat -c 2 -D hw:0,0 test.wav`
then `arecord` does not record any noise

4.
if the mic is working just before issuing the command
`pasuspender -- yes > /dev/null`
and I suspend and resume without killing the process
then the mic is *not* working after the resume, nor after
I kill the process after resuming

On 18 April 2014 16:39, Felipe Sateler <fsateler at debian.org> wrote:
> On Fri, Apr 18, 2014 at 1:31 PM, Stefan Krastanov
> <krastanov.stefan at gmail.com> wrote:
>> Hi,
>>
>> Here is the new log file:
>>
>> - starting pulse audio (not working)
>> - tapping on the internal mic (not working)
>> - tapping *very* hard on the internal mic (starts working)
>> - suspend and resume (stops working)
>> - starting gnome-sound-recorder (not working)
>> - jiggling the volume bar in the control panel (starts working)
>
> This is starting to look like a kernel bug. Apparently the device is
> put into suspended mode and for some reason it cannot bring it back:
>
>> (  22.139|  16.978) I: [alsa-source-CS4206 Analog] alsa-source.c: Scheduling delay of 3955.27 ms > 10.00 ms, you might want
>> (  22.139|   0.000) D: [alsa-source-CS4206 Analog] alsa-util.c: Got POLLERR from ALSA
>> (  22.139|   0.000) D: [alsa-source-CS4206 Analog] alsa-util.c: Got POLLIN from ALSA
>> (  22.139|   0.000) D: [alsa-source-CS4206 Analog] alsa-util.c: PCM state is SUSPENDED
>> (  22.139|   0.000) I: [alsa-source-CS4206 Analog] (alsa-lib)pcm_hw.c: SNDRV_PCM_IOCTL_RESUME failed (-38)
>> (  22.144|   0.004) I: [alsa-source-CS4206 Analog] alsa-source.c: Starting capture.
>
> Lets try again but now bypassing pulseaudio:
>
> pasuspender -- arecord -vv -fdat -c 2 -D hw:0,0 test.wav
>
> I had to specify the -c 2 argument for the number of channels, but you
> might not need to.
>
> Try this after the mic died with pulseaudio, as the dying might be
> caused by pulseaudio sending the device to sleep.
>
> also maybe you could try doing:
>
> pasuspender -- yes > /dev/null
>
> This will suspend pulseaudio until you interrupt that process. Then
> suspend the computer, bring it back and then try to record using
> arecord. These two tests should tell us if the problem is in ALSA
> rather than in pa.
>
>>
>> If it might help, I cat repeat this with a stopwatch for easier
>> reading of the log file?
>
> I didn't think of that! I'm going to ask that next time :)
>
>>
>> Thanks for looking into this.
>
> No worries
>
>
> --
>
> Saludos,
> Felipe Sateler