On Tue, 2008-03-11 at 15:24 -0600, Bjorn Helgaas wrote:
I cannot judge about the "cleanness" because of possible changes of
initialization time of pci/pnp or whatever.
I can confirm that ACPI resource declarations (especially (only?) in the
general motherboard devices) are often rather broken.
Therefore I like your general approach compared to the per machine dmi
match quirks. I expect the latter (dmi matching) is rather hopeless.
Thomas
--