Re: [RFC] Fix Device Power Management States

[David Brownell <david-b@pacbell.net>]

On Monday 09 August 2004 9:55 pm, Patrick Mochel wrote:
> 
> On Tue, 10 Aug 2004, Benjamin Herrenschmidt wrote:
> 
> 
> > What about partial tree ? We need to suspend childs first and we need to
> > tied PM transition with dev_start/stop (or have some way to indicate the
> > device we want it to auto-resume when it gets a request, or something).
> > We need to work out policy a bit more here I suppose...
> 
> Policy can come later; we have to have a working model first.

Suspending a partial tree is part of the "device power management"
problem ... it's not a policy, and deferring it would punt on one of
the more basic problems.

A policy would be IMO more like "how far should I suspend this",
"can this device wake out of its suspend state", or more interestingly
"how should these devices interact".

So:  "suspend everything using the 48 MHz clock, so we can
enter a deeper system sleep state" would be a kind of policy.
At least so long as those clock relationships don't appear
explicitly in the driver model ... they could be recorded in
device-specific power state descriptors.
 
- Dave