Re: [RFC] Fix Device Power Management States

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

On Tuesday 10 August 2004 3:13 am, Pavel Machek wrote:
> Hi!
> 
> > > Well, "no DMA" needs to be part of definition, too, because some
> > > devices (USB) do DMA only if they have nothing to do.

I think that should read "even if they have ...", not "only if ...".


> > I don't understand; that doesn't sound healthy.
> 
> It is not healthy. It is basicaly misdesigned piece of hardware called
> UHCI. It simply does DMA all the time :-(.

Unless it's suspended ... and I confess I've only had time to
make sure that OHCI and EHCI suspend properly using the
newish CONFIG_USB_SUSPEND patch.  UHCI is different
from those other mainstream controllers, it doesn't split
out its periodic schedule processing.

Keep in mind that to properly quiesce a USB controller, you've
got to quiesce every driver for every device hooked up to
that USB bus.  There's no escaping the bottom-up suspend
or top-down-resume processes, which makes me wonder
how Patrick's proposed patch can work for it...

- Dave