Re: Power Management Framework architectural design

[Martin Husemann <martin%duskware.de@localhost>]

On Tue, Apr 15, 2008 at 09:52:28AM -0700, Ted Lemon wrote:
> on even when the CPU is off - the wireless device.   It remains on and  
> running at very low power in order to maintain the mesh network.   So  
> it would be wrong to power it off when powering off whatever bus it's  
> on.

Yeah, this is a central point of the discussion: pmf can (currently) only do
tree walks and power down things bottom up.

Now what you (and others) describe makes sense for obio/mainbus like "buses",
but do we expect, say, a pci card to survive the pci bus bridge being powered
down? For what value of "powered down"? And can we make pmf deal with this?

Martin