Bug#637392: Fwd: linux-image-2.6.32-5-amd64: [drm:r600_audio_bits_per_sample] *ERROR* Unknown bits per sample 0xf using 16 instead

[Volodymyr Shcherbyna <volodymyr@shcherbyna.com>]

I decided to repeat the case and I have started machine without any special software, i.e., no qtcreator, no vlc, just BOINC running, and I have the problem after 3 hours of idle running. It seems like my dmesg is full of 

"[drm:r600_audio_bits_per_sample] *ERROR* Unknown bits per sample 0xf using 16 instead."

It is hard to say when exactly it happened, as dmesg is full of these lines, all previous history is not saved.

Thanks,

-- 
with best regards, Volodymyr

---------- Forwarded message ----------
From: Volodymyr Shcherbyna <volodymyr@shcherbyna.com>
Date: 2011/8/11
Subject: Re: linux-image-2.6.32-5-amd64: [drm:r600_audio_bits_per_sample] *ERROR* Unknown bits per sample 0xf using 16 instead
To: Jonathan Nieder <jrnieder@gmail.com>
Cc: 637392@bugs.debian.org


Hello Jonathan,

Unfortunately, I don't have many details about this issue yet. It started to happen only after 5th of August, and I don't remember any major updates near that date. There was only one update which updated some smb packages, IIRC.

Issue usually happens after I boot machine and work for some time. Twice it happened when I was coding in qtcreator, maybe this is related.  I also experienced similar problem when booting - but I thought it was a kernel panic related to my webcam (webcam is not new, it is attached for my machine for 2 years), and I don't have stack trace of that case.

In order to deduct some trigger I guess it is a combination of high CPU (I donate to BOINC), qtcreator and vnc which is playing podcasts ... I can't repro the issue however ...

Will try soon sid and upstream kernel, I had before 3.0 rc7 and it worked well right away.

Thank you,

--
with best regards, Volodymyr

2011/8/10 Jonathan Nieder <jrnieder@gmail.com>

Hi Volodymyr,

Volodymyr Shcherbyna wrote:

> Hello, it seems like my video driver is failing. My screen just
> switches off itself and I see the following messages in my dmesg:
[...]
> The last line in log is [drm:r600_audio_bits_per_sample] *ERROR*
> Unknown bits per sample 0xf using 16 instead and I see a lot of them
> ... The only way to recover from this situation is to reboot
> machine.

Hmm, that's no good.  Does this always happen at boot, or after a
certain trigger, or at seemingly random times?  Is it easy to
reproduce?  Did it start happening with an upgrade, or has it always
happened on this machine?  Does it happen with the kernel from sid,
too?  Any other weird symptoms?

This might have been fixed by v2.6.33-rc8~14^2~1^2~6 = v2.6.32.0.0~67
(drm/radeon/kms: suspend and resume audio stuff, 2010-01-28), so
results from testing the kernel in sid would be very helpful.

Thanks for reporting.

Regards,
Jonathan