Suggestion for underlying design of power settings infrastructure (was: Re: Power Management and OpenMoko - third power setting)

[Liam Girdwood]

On Thu, 2008-02-28 at 11:12 +0100, Michael 'Mickey' Lauer wrote:
> On Thursday 28 February 2008 01:26:04 Andy Green wrote:
> > The real one big win that we have on our side is the full suspend, that
> > is not granular at all because it turns off the CPU core then.  I think
> > instead of exposing a whole 1950s style control panel of options it
> > means we just need to keep things turned off until we need them and as
> > soon as we are done with them, and keep the whole phone in suspend the
> > vast bulk of the time.  All the other options won't make any real
> > difference I think because I just look at the backlight consumption and
> > the rest is largely insignificant.
> 
> Agreed. If we use a clever policy system, there's almost no use for manual 
> intervention for powering on and off subsystems.
> 
> Speaking about that, how is userland supposed to control the Wifi module and 
> the USB subsystem? Do we need to unload/load modules to control or can we add 
> a power control sysfs node like we have for BT, GPS, GSM ?
> 

The voltage regulator API may help :-

http://opensource.wolfsonmicro.com/node/15

This will allow drivers to power down when not used. e.g. Wifi could
power down with "ifconfig down" 

Liam