[alsa-devel] [RFC_i/iv 1/3] ASoC: Decouple DAPM from CODECs. Part core (will be squashed)

[Mark Brown]

On Sat, Oct 30, 2010 at 07:01:54PM +0100, Liam Girdwood wrote:
> On Fri, 2010-10-29 at 13:04 -0700, Mark Brown wrote:

> > >  		run_delayed_work(&card->rtd[i].delayed_work);
> > > -		card->rtd[i].codec->suspend_bias_level = card->rtd[i].codec->bias_level;
> > > +		card->rtd[i].codec->dapm->suspend_bias_level = card->rtd[i].codec->dapm->bias_level;

> Btw, not sure what you mean by this 1 line change being run multiple
> times being different to the old line.

It's not different, a lot of the stuff here is things that we need to do
in the series that are thrown up by the series as a result of it
touching all relevant places.  I tried to mention this in the places I
noted it, but some of them I will have forgotten.

> > I'd be more inclined to just embed the struct in the objects that need
> > it rather than individually allocating them - it saves error checking
> > and deallocation, and I can't see any cases where we'd want to
> > optionally have a DAPM object.

> Yeah, the reason for this is because it was unclear at the time whether
> we will need to share a context between DAI backend and frontends. So
> this can be embedded now.

Jarkko, with this change I think this first patch is good to go - I
think we should try to get the early bits in while we're iterating on
the rest.