[alsa-devel] Possible fix for snd-hda-intel model=no-jd failing since ~linux-3.9-rc1

[Adam Richter]

On Thu, 9/11/14, Takashi Iwai <tiwai at suse.de> wrote:

> Could you give alsa-info.sh output on your machine?  Otherwise it's difficult to analyze.

OK.  I have attached it.  In a few places, I replaced text with "[DELETED]" to avoid identifying the product.

> Speaking of your patch: add_jack_kctl() itself has a check of is_jack_detectable() for phantom jacks,
> so basically this shouldn'tmake any difference.  So we need to check more deeply why this
> change is really needed.


Right. This is some kind of initialization order problem where the following data flow fails: codec->no_jack_detect ==> jack->phantom_jack ==> jack->pin_sense.  The Intel driver has a lot of "codec" sub-drivers, and I wasn't sure if would would really be practical for me to understand and get the relevant codec maintainers to approve if necessary.  The fix I proposed is something I see as a temporary very low overhead workaround (only adds a few instructions, and only at initialization) until a correct fix is implemented.  However, if this can be fixed properly quicly enoughtso avoid the need for a temporary workaround, that would be great. 
 
>thanks,
>
>Takashi
 
Thank you for your analysis of this.

Adam
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: alsa-info.txt
URL: <http://mailman.alsa-project.org/pipermail/alsa-devel/attachments/20140911/6da8e70d/attachment-0001.txt>