On Fri, Oct 15, 2010 at 08:26:34AM +0300, Jukka Ruohonen wrote: > The task is not trivial. On modern x86, practically *everything* that > attachs has an ACPI counterpart. In a way we are thinking this backwards: > the attachment should perhaps be done via ACPI that has information about > the "natural" device tree (I recommend to boot with ACPIVERBOSE option and > observe the output). This is how it is supposedly done in Windows. And > consequently, *most* (MI) drivers that work on x86 need to eventually call > (MD) ACPI callbacks, and vice versa. Bringing this all together in a clean > (MI) implementation is hard and requires substantial changes, to say the > least.
As an addition, due reasons stated above, I object anything that tries to make a case for a single driver from acpi(4) -- be it acpivga(4), acpicpu(4), or the ISA and PCI cases discussed previously. This should be solved once and for all, for all acpi(4) and for all pci(4), isa(4), ... Otherwise we end up with god-awful mess. If such a solution comes to existence, we are happy to refactor acpi(4). During the ten years that ACPI has been in NetBSD, several people have tried a solution without much success. I have personally tried twice, and failed already at the self-criticism stage. - Jukka.