Bjorn, I like how pci_fixup_device() invokes the quirk when we want at pci_enable_device() time w/o cluttering the code with VIA specific hooks.
I think you've also identified an improvement by checking acpi_irq_model -- as the PCI config space IRQ register is defined only in PIC-mode -- so one must assume that the quirky via HW can't depend on us writing reserved bits for IRQs > 15. But checking skip_ioapic_setup in the non-ACPI case isn't quite right. This is set for "noapic". But it is not set in the PIC-mode case where the kernel supports IOAPIC but the hardware does not -- in that case the quirk would erroneously exit. Also, the original quirk_via_irqpic() had a udelay(15) before the write -- I have no idea if that was significant or not -- maybe soembody else on the list does -- as none of us have VIA documentation... thanks, -Len ps. we need to fix this on 2.4 also. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/