[alsa-devel] [PATCH] hda: STAC927x power down inactive DACs

[Takashi Iwai]

At Wed, 30 Jan 2008 17:33:57 -0500,
Matthew Ranostay wrote:
> 
> Takashi Iwai wrote:
> > At Wed, 30 Jan 2008 09:14:02 -0500,
> > Matthew Ranostay wrote:
> >> On several laptops that have STAC9228 codecs have unused DACs, this powers them down to a D3 state.
> >>
> >> Signed-off-by: Matthew Ranostay <mranostay at embeddedalley.com>
> > 
> > I think they will be eventually powered up again in
> > hda_set_power_state() in hda_codec.c.  As a temporary solution, we can
> > have a sort of blacklist in struct hda_codec for always turning the
> > power off for such widgets.
> > 
> > As a long-term solution, the automatic detection of such widgets would
> > be nice...
> > 
> > 
> > Takashi
> > 
> > 
> >> ---
> >> diff -r 86c8c0a82bb9 pci/hda/patch_sigmatel.c
> >> --- a/pci/hda/patch_sigmatel.c	Wed Jan 30 11:52:33 2008 +0100
> >> +++ b/pci/hda/patch_sigmatel.c	Wed Jan 30 08:58:37 2008 -0500
> >> @@ -288,6 +288,10 @@ static hda_nid_t stac927x_adc_nids[3] = 
> >>  
> >>  static hda_nid_t stac927x_mux_nids[3] = {
> >>          0x15, 0x16, 0x17
> >> +};
> >> +
> >> +static hda_nid_t stac927x_dac_nids[5] = {
> >> +	0x02, 0x03, 0x04, 0x05, 0x06
> >>  };
> >>  
> >>  static hda_nid_t stac927x_dmux_nids[1] = {
> >> @@ -3537,7 +3541,7 @@ static int patch_stac927x(struct hda_cod
> >>  static int patch_stac927x(struct hda_codec *codec)
> >>  {
> >>  	struct sigmatel_spec *spec;
> >> -	int err;
> >> +	int err, i;
> >>  
> >>  	spec  = kzalloc(sizeof(*spec), GFP_KERNEL);
> >>  	if (spec == NULL)
> >> @@ -3628,6 +3632,14 @@ static int patch_stac927x(struct hda_cod
> >>  		stac92xx_free(codec);
> >>  		return err;
> >>  	}
> >> +
> >> +	/* on some 3-stack systems not all the DACs are used
> >> + 	 * this powers down ones that aren't used
> >> + 	 */
> >> +	for (i = 0; i < ARRAY_SIZE(stac927x_dac_nids); i++)
> >> +		if (!is_in_dac_nids(spec, stac927x_dac_nids[i]))
> >> +			snd_hda_codec_write_cache(codec, stac927x_dac_nids[i],
> >> +				0, AC_VERB_SET_POWER_STATE, AC_PWRST_D3);
> >>  
> >>  	codec->patch_ops = stac92xx_patch_ops;
> >>
> > 
> 
> I think this what you had in mind.
> Also is putting the function call in stac92xx_resume correct?

Yeah, thanks, I forgot it.
I applied it to my tree.


Takashi