[alsa-devel] [PATCH] ALSA: hda - Fix Dock Headphone on Thinkpad X250 seen as a Line Out

[Takashi Iwai]

At Wed, 24 Jun 2015 13:06:05 +0200,
David Henningsson wrote:
> 
> 
> 
> On 2015-06-24 11:00, Takashi Iwai wrote:
> > At Wed, 24 Jun 2015 10:46:33 +0200,
> > David Henningsson wrote:
> >>
> >> Thinkpad X250, when attached to a dock, has two headphone outs but
> >> no line out. Make sure we don't try to turn this into one headphone
> >> and one line out (since that disables the headphone amp on the dock).
> >>
> >> Alsa-info at http://www.alsa-project.org/db/?f=36f8764e1d782397928feec715d0ef90dfddd4c1
> >>
> >> Cc: stable at vger.kernel.org
> >> Signed-off-by: David Henningsson <david.henningsson at canonical.com>
> >
> > This is one of fixes I had in mind (and asked for testing in bugzilla
> > kernel.org report).
> 
> Sorry, I'm missing context. What report?
> 
> > One drawback by this is that the auto-mute
> > *between* the headphones won't work like before.  If one has expected
> > it, then this can be seen as a regression.
> 
> Fair enough, we can drop stable from the patch since this could be seen 
> as behavioural change rather than bug fix.
> 
> > Another fix was to change this as a proper line-out.  This will keep
> > the current auto-mute behavior.
> >
> > OTOH, dealing both jacks equally is also not bad, per se.
> 
> The tricky thing about routing is that everyone wants it different, so 
> pleasing everyone seems to be an impossible equation :-/
> 
> I would choose the default behaviour (headphone or line out) based on 
> the icon on the jack. I should double-check that with someone who has 
> the hardware. And then, if someone wants another behaviour then 
> hdajackretask can be used. Or we can add a jack mode kcontrol for that 
> jack, if you prefer, although I'd have to check how PulseAudio handles 
> that (not that well, is my gut feeling).

I have no preference in this regard.  My only slight concern was the
behavior change by this patch.  It has both sides, one can see as an
improvement while another as a regression.

Currently I'm inclined to take this patch as is, since this would make
things working in a little obstacle.  Once when we get a real
regression report due to the lack of exclusive mute control, we may
provide another model option as a compromise.

But, I'm open about this, really.  So, if anyone has a good argument,
let me know.


thanks,

Takashi