Re: acpivga(4) v. MI display controls

[Grégoire Sutre <gregoire.sutre%gmail.com@localhost>]

On 10/20/2010 09:01, Jukka Ruohonen wrote:

> The above example also reveals the devices (in this machine) that
> reference the ACPI embedded controller's operation regions. Thus, the
> three children should be attached under acpiec(4), or more
> conservatively, these should at least never be attached before
> acpiec(4).

Currently, most ACPI devices attach directly under acpi(4).  I believe that
it would make sense, at least conceptually, to follow the ACPI tree in auto-
configuration of acpi(4) devices.  This way, the acpi(4) attachment tree and
the ACPI tree could be synchronized.  But I'm afraid this would not help for
the "power resource" problem you mentioned.

Still, for acpiec(4), this would not be sufficient since some dependencies
are not explicit in the ACPI tree: some devices (e.g. acpivga(4)) should
attach after acpiec(4) even though they are not descendants (in the ACPI
tree) of the ACPI embedded controller.

Grégoire