snd_hda_intel headphone output disabling upon complete insertion of plug

snd_hda_intel headphone output disabling upon complete insertion of plug

[lkml]

Hello list,

Somewhere between 3.11.0-rc7-gfa8218d and 3.12.0-rc4-g8b5ede6 this
broke.

The machine is an x61s thinkpad, with the headphone output unmuted and
headphones partially inserted sound is audible.  If the output is muted,
the output shuts off.  With the output unmuted, upon inserting the
headphone plug fully the output shuts off following a blink of a delay.
If I slowly remove the headphone plug returning it to a partially inserted
state, the audio returns after another delay.

It seems like a bug in the headphone amplifier power management, as if
it's doing the inverse of what it needs to do.  I'll try find time to
bisect, just wanted to get this out there though in case any of you
recognize the problem and know what might have caused it.

Regards,
Vito Caputo

Re: snd_hda_intel headphone output disabling upon complete insertion of plug

[Mark Knecht]

On Tue, Oct 8, 2013 at 2:07 PM,  <lkml@pengaru.com> wrote:
> Hello list,
>
> Somewhere between 3.11.0-rc7-gfa8218d and 3.12.0-rc4-g8b5ede6 this
> broke.
>
> The machine is an x61s thinkpad, with the headphone output unmuted and
> headphones partially inserted sound is audible.  If the output is muted,
> the output shuts off.  With the output unmuted, upon inserting the
> headphone plug fully the output shuts off following a blink of a delay.
> If I slowly remove the headphone plug returning it to a partially inserted
> state, the audio returns after another delay.
>
> It seems like a bug in the headphone amplifier power management, as if
> it's doing the inverse of what it needs to do.  I'll try find time to
> bisect, just wanted to get this out there though in case any of you
> recognize the problem and know what might have caused it.
>
> Regards,
> Vito Caputo

Sometimes across kernel revisions I've had problems like this.
(Although not that specific one!) It seems that the asound.state file
format changes and alsa renumbers their internal controls. You might
try saving the existing file elsewhere and then setting up alsa again
from scratch. That solved my last problem of that type.

HTH,
Mark

Re: snd_hda_intel headphone output disabling upon complete insertion of plug

[lkml]

On Tue, Oct 08, 2013 at 02:34:33PM -0700, Mark Knecht wrote:
> On Tue, Oct 8, 2013 at 2:07 PM,  <lkml@pengaru.com> wrote:
> > Hello list,
> >
> > Somewhere between 3.11.0-rc7-gfa8218d and 3.12.0-rc4-g8b5ede6 this
> > broke.
> >
> > The machine is an x61s thinkpad, with the headphone output unmuted and
> > headphones partially inserted sound is audible.  If the output is muted,
> > the output shuts off.  With the output unmuted, upon inserting the
> > headphone plug fully the output shuts off following a blink of a delay.
> > If I slowly remove the headphone plug returning it to a partially inserted
> > state, the audio returns after another delay.
> >
> > It seems like a bug in the headphone amplifier power management, as if
> > it's doing the inverse of what it needs to do.  I'll try find time to
> > bisect, just wanted to get this out there though in case any of you
> > recognize the problem and know what might have caused it.
> >
> > Regards,
> > Vito Caputo
> 
> Sometimes across kernel revisions I've had problems like this.
> (Although not that specific one!) It seems that the asound.state file
> format changes and alsa renumbers their internal controls. You might
> try saving the existing file elsewhere and then setting up alsa again
> from scratch. That solved my last problem of that type.
> 

NAK, has no effect.  Looking at the diff between the revisions under
sound/pci/hda there appears to be quite a bit of activity.  Hopefully I
can get around to bisecting this weekend.  Leaving the headphones 90%
inserted is workable for now.

Regards,
Vito Caputo