[alsa-devel] HD-audio generic parser merged to linux-next

Takashi Iwai tiwai at suse.de
Thu Jan 24 07:31:09 CET 2013

At Thu, 24 Jan 2013 11:26:35 +0800,
Raymond Yau wrote:
> >
> > > Some feature is only available when the hardware meet minimum
> requirement
> > > e.g. independent headphone require at least two DAC
> >
> > Why it's a regression?
> >
> Those hda codecs  with a single DAC (e.g. stac920x, alc260) and ad1986a
> does no support independent headphone

I still don't understand what you are trying to say...
The independent HP mode is created only for certain codecs like AD and
VT codecs, so far.

Moreover, how would "independent" HP work with a single DAC at all?
The independent HP means that the individual PCM stream can be played
through the headphone.  It doesn't make any sense for the hardware
with a single DAC.

> Also those alc66x datasheet does not mention the support of multi-streaming
> playback

There has been no change wrt Realtek codec.
We're talking about regression.

> Don't create the control when there is internal speaker

More context please.

> When driver assign two DAC to internal speaker and subwoofer (vt1802 on
> Asus g75) and it is confusing when there is no headphone playback volume
> control

It's rather a fix.  The headphone volume was wrongly picked up
although it's identical with the speaker volume.

> Don't create the control when there is no headphone Jack  when the user is
> retasked  the headphone Jack to Mic by early patching

When user says it's a mic jack, how can you know it's a headphone?

> Automatically disable for those headphone mic  jack when the HP Jack is
> retasked as mic

The old behavior was rather buggy, and this is actually a fix.

Think of the other scenarios: if we inactivate the control element,
you may get again the crash of application as you've seen with the old
AD codec parser.  OTOH, if we change the value of automute mode enum,
how should we do once when the independent HP is turned off again?
You can't know whether user want to keep the value off or not.


More information about the Alsa-devel mailing list