jmcneill@invisible.ca said:
>   3. Eventually, nuke the powerhook_establish/powerhook_disestablish
>      framework

Great - this doesn't deal with hierarchical buses and needs to die.

> device class policies along with a sample power 
> management handler for the wskbd(4) driver (input class), azalia(4) (audio 
> class), and vesafb(4) (display class).
> [...]
> Audio devices are transitioned to D0

As I understand, you are using the ACPI nomenclature for device
states here. That's reasonable because it comprises some common
ground, but the namespace used should be made obvious everywhere.

For user-visible actions, entities like up/suspended/hibernated
should be used which are translated to ACPI states where ACPI is
in control, and to bus/device specific states otherwise.

Power management of eg PCI(express) and USB lives in different
namespaces. These use the same letters to designate states, but
there are nontrivial dependencies eg between power states of buses
and devices.
We should obey the namespace everywhere and translate explicitely
where necessary. Otherwise we'd get a lot of ambiguities to
be solved at the driver level.

best regards
Matthias


Forschungszentrum Juelich GmbH
52425 Juelich

Sitz der Gesellschaft: Juelich
Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498
Vorsitzende des Aufsichtsrats: MinDirig'in Baerbel Brumme-Bothe
Vorstand: Prof. Dr. Achim Bachem (Vorsitzender), Dr. Ulrich Krafft (stellv. 
Vorsitzender)