Re: [PATCH 7/8] ASoC: tlv320aic3x: add initial usage of regulator framework to control avdd_dac

[Eero Nurkkala <ext-eero.nurkkala@nokia.com>]

On Thu, 2009-10-08 at 18:01 +0200, ext Mark Brown wrote:
> On 8 Oct 2009, at 16:44, <ext-Eero.Nurkkala@nokia.com> wrote:
> 
> > Mark Brown wrote:
> >>
> >>>                                         Also, this is regulator  
> >>> thing
> >>> is highly platform dependent, not aic3x related really at all, so is
> >>> this the correct place... Just a thought, dont take it too  
> >>> seriously ;)
> >>
> >> I'm not sure what you mean by this?
> >
> > You may power the aic3x from a fixed source, or from multiple  
> > sources, with
> > and without any regulator in between. It's up to the HW and HW design.
> 
> The regulator API can cope with all this pretty transparently - if  
> multiple supplies come from the same regulator the API will hide that  
> from the consumer. There is a fixed voltage regulator driver which can  
> be used to represent supplies with no soft control.
> 
> > Moreover, you don't _power off_ (turn the regulator off) the analog  
> > voltages
> > of aic3x; things won't work. So it's not like a switch everybody may  
> > use. Or
> > nothing prevent you from experiencing that...
> 
> I'd expect the usage would be that after the audio subsystem has been  
> idle for some configurable period of time the core would bring the  
> audio subsystem down to bias off, at which point supplies could also  
> be switched off.

Right. That would also sound like the RST line also needs also be
asserted, and then rewriting all register contents upon wakeup? And also
redirecting all i2c traffic to the cache instead of any real i2c writes
(meanwhile the device is shut down)? Like in tpa6130?

- Eero