[alsa-devel] Add AZX_DCAPS_I915_POWEREWELL on SKL is OK?

[Yang, Libin]

Hi Takashi,


> -----Original Message-----
> From: Takashi Iwai [mailto:tiwai at suse.de]
> Sent: Thursday, March 26, 2015 2:29 PM
> To: Yang, Libin
> Cc: alsa-devel at alsa-project.org
> Subject: Re: Add AZX_DCAPS_I915_POWEREWELL on SKL is OK?
> 
> At Thu, 26 Mar 2015 05:40:20 +0000,
> Yang, Libin wrote:
> >
> > Hi Takashi,
> >
> > Is it OK to add AZX_DCAPS_I915_POWEREWELL for SKL in hda_intel.c?
> >
> > SKL HDA controller is not in powerwell domain. But its
> > HDMI codec is in the domain. My understanding is if the flag
> > is added here, it means the controller is in the domain. But I can't
> > find a good place to add a flag to indicate the HDMI codec need
> > the power domain.
> 
> The question is when you need the powerwell enable/disable.
> If the powerwell has to be adjusted before *any* HDMI codec access,
> we'd need this even at probing the codec.  Currently HD-audio
> controller driver probes a codec by reading VENDOR_ID parameter of
> the
> root node.  This was supposed to work even at codec D3, except for
> Intel HDMI controllers.
> 
> Now, HDMI codec is on the common bus on SKL.  Does this mean that
> reading the root node parameter works even without powerwell?
> This has to be clarified at first.  I asked this in the last meeting,
> but it seems still unclear.
> 
> If the root node parameter can be read without powerwell, then the
> controller doesn't need it.  Instead, a similar code needs to be added
> in the HDMI codec side.  OTOH, if the root parameter reads already
> requires the powerwell, then yes, we need to add the DCAPS to the
> controller.
> 


Thanks. I will confirm with our internal silicon team for your question.

> 
> Takashi