Re: [RFC] dynamic device power management proposal

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

On Monday 19 March 2007 8:44 am, Alan Stern wrote:
> On Mon, 19 Mar 2007, Shaohua Li wrote:
> 
> > ...
> > Basically we need device driver support, a kernel framework and a policy
> > (determine when to change a device’s power state).
> 
> A lot of development along these lines has already been going on in the 
> USB subsystem.  It isn't complete yet, but a lot of the ideas you raise 
> have already been implemented.

ISTR pointing out a few years ago that USB makes a good testbed for
such things, in technical terms ... it's got the "complete problem"
wrapped up in one more-or-less modern subsystem.  So making it do
PM well has been an object lesson in how the Linux PM infrastructure
works -- or, too often, doesn't.

The most troublesome spot being that ACPI interferes on PCs, so we
can't yet do stuff like put EHCI/OHCI/UHCI controllers into PCI_D2 at
runtime and rely on remote wakeup.   But embedded Linux systems make
it easy to work with saner frameworks; on those systems you can see
even more of the pieces working well together.  :)

- Dave