Kernel 2.6.33: PXA270 Audio does not resume after resume from suspend state

Kernel 2.6.33: PXA270 Audio does not resume after resume from suspend state

[Shane Volpe]

I have upgraded an embedded system (PXA270) to the 2.6.33 kernel (from
2.6.23).  Now when I run an audio program (ogg123) and put the system
into standby then wake the system up the audio does not work (continue
playing) after resume until I terminate ogg123 and start again.  After
restarting ogg123 the audio works fine with out reseting the system or
changing anything else on the system.

The system is using the UCB1400 AC97 codec and just as added
information when the audio is locked up, after resume, the touchscreen
on the system still works which is also controlled by the UCB1400
audio codec.  I have also tried several other audio players and the
results are the same.   Has anyone else seen this bug and does anyone
have suggestions on further tests to perform to narrow down were the
bug may reside?
Regards,
Shane

Kernel 2.6.33: PXA270 Audio does not resume after resume from suspend state

[Daniel Mack]

On Tue, Jun 08, 2010 at 10:38:36AM -0400, Shane Volpe wrote:
> I have upgraded an embedded system (PXA270) to the 2.6.33 kernel (from
> 2.6.23).  Now when I run an audio program (ogg123) and put the system
> into standby then wake the system up the audio does not work (continue
> playing) after resume until I terminate ogg123 and start again.  After
> restarting ogg123 the audio works fine with out reseting the system or
> changing anything else on the system.
> 
> The system is using the UCB1400 AC97 codec and just as added
> information when the audio is locked up, after resume, the touchscreen
> on the system still works which is also controlled by the UCB1400
> audio codec.  I have also tried several other audio players and the
> results are the same.   Has anyone else seen this bug and does anyone
> have suggestions on further tests to perform to narrow down were the
> bug may reside?

I've had similar issues some time ago. However, we used I2S over SSP for
our system, not AC97, and suspend/resume works fine now.

You may have a look at commit 026384d614b ("ASoC: fix PXA SSP port
resume") and check whether something similar must be done for the AC97
controller part.

Daniel