linux-next - strange audio bug

linux-next - strange audio bug

[Valdis Kletnieks]

[-- Attachment #1.1: Type: text/plain, Size: 800 bytes --]

Next-20150310 works OK, next-20150324 has the bug.

Dell Latitude E6530 laptop, lspci tells me:

01:00.1 Audio device: NVIDIA Corporation GF108 High Definition Audio Controller (rev a1)
        Subsystem: Dell Device 0535
        ...
        Kernel driver in use: snd_hda_intel

Symptoms:  Audio works perfectly at boot, and for some amount
of time after.  But somewhere around 6-8 hours of uptime, it gets
into a strange state where I'll get no audio... and then it will
cut in for all of 2 seconds or so (fairly consistent, that part),
and then cut back out for anywhere from 10 seconds to a minute,
then another 2 seconds of sound.. lather rinse repeat.

Before I start the thankless task of bisecting this (which means I'll
be able to do only 1, *maybe* 2 steps a day), does this ring any bells?

[-- Attachment #1.2: Type: application/pgp-signature, Size: 848 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: linux-next - strange audio bug

[Hui Wang]

On 04/03/2015 08:34 AM, Valdis Kletnieks wrote:
> Next-20150310 works OK, next-20150324 has the bug.
>
> Dell Latitude E6530 laptop, lspci tells me:
>
> 01:00.1 Audio device: NVIDIA Corporation GF108 High Definition Audio Controller (rev a1)
>          Subsystem: Dell Device 0535
>          ...
>          Kernel driver in use: snd_hda_intel
>
> Symptoms:  Audio works perfectly at boot, and for some amount
> of time after.  But somewhere around 6-8 hours of uptime, it gets
> into a strange state where I'll get no audio... and then it will
It sounds like the suspend or resume of power management introduces this 
problem. Probably you could reproduce this problem very soon just 
repeatedly suspend and resume the system.
> cut in for all of 2 seconds or so (fairly consistent, that part),
> and then cut back out for anywhere from 10 seconds to a minute,
> then another 2 seconds of sound.. lather rinse repeat.
>
> Before I start the thankless task of bisecting this (which means I'll
> be able to do only 1, *maybe* 2 steps a day), does this ring any bells?
>
>
> _______________________________________________
> Alsa-devel mailing list
> Alsa-devel@alsa-project.org
> http://mailman.alsa-project.org/mailman/listinfo/alsa-devel

Re: [alsa-devel] linux-next - strange audio bug

[Valdis.Kletnieks]

[-- Attachment #1: Type: text/plain, Size: 885 bytes --]

On Fri, 03 Apr 2015 08:44:12 +0800, Hui Wang said:

> It sounds like the suspend or resume of power management introduces this
> problem. Probably you could reproduce this problem very soon just
> repeatedly suspend and resume the system.

I don't do suspend on this laptop, I poweroff and reboot when I get
wherever I'm going.

However... device-level power management? Hmm..  I *do* remember that
each time I found it dead, I was coming back to my laptop from doing something
else, so the sound would have been idle....

commit c4c2533f802d6877803c4d778def43d8a122f27b
Author: Takashi Iwai <tiwai@suse.de>
Date:   Tue Mar 3 17:22:12 2015 +0100

    ALSA: hda - Fix possible runtime PM refcount unbalance

That landed in next-20150324 - which is broken for me.

Looks worth trying a revert before I bisect, thanks for the hint, will
report back (probably tomorrow) on what happens...

[-- Attachment #2: Type: application/pgp-signature, Size: 848 bytes --]

Re: linux-next - strange audio bug

[Takashi Iwai]

At Thu, 02 Apr 2015 20:34:35 -0400,
Valdis Kletnieks wrote:
> 
> Next-20150310 works OK, next-20150324 has the bug.
> 
> Dell Latitude E6530 laptop, lspci tells me:
> 
> 01:00.1 Audio device: NVIDIA Corporation GF108 High Definition Audio Controller (rev a1)
>         Subsystem: Dell Device 0535
>         ...
>         Kernel driver in use: snd_hda_intel
> 
> Symptoms:  Audio works perfectly at boot, and for some amount
> of time after.  But somewhere around 6-8 hours of uptime, it gets
> into a strange state where I'll get no audio... and then it will
> cut in for all of 2 seconds or so (fairly consistent, that part),
> and then cut back out for anywhere from 10 seconds to a minute,
> then another 2 seconds of sound.. lather rinse repeat.
> 
> Before I start the thankless task of bisecting this (which means I'll
> be able to do only 1, *maybe* 2 steps a day), does this ring any bells?

Try for-next branch of sound git tree
   git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git

Now I merged fixes that might be relevant with the runtime PM on some
codecs, so it might work better.

In anyway, please give alsa-info.sh outputs at good and bad-working
states.


thanks,

Takashi