Re: [ACPI] IRQ problems with 20020611

[Arndt S. <ab...@sr...>]

Hi Johnathan,

I don't know if this also applies to your specific system, but if
something goes wrong with regard to ACPI and interrupts, it is usually
a good idea to check Dominik Brodowski's pciirq patches from

    http://www.brodo.de/english/pub/acpi/pci_irq/index.html

If it still doesn't work with Dominik's patch, I suggest you write to
him directly at de...@br.... If it does work, please also inform
the list so that Andy and Paul can submit the fix for inclusion into
the Linux 2.5 kernel.

Good luck,
Arndt

On Thu, Jun 20, 2002 at 10:00:22PM -0400, Johnathan Hicks wrote:
> Hi.
> 
> I'm having IRQ problems with 20020611 on 2.4.18 and the 2.4.19-pre10 what I
> massaged the patch into. I did not have trouble with 20020517 on the same
> kernel versions. I looked at the kernel messages and saw a bunch "PCI BIOS
> passed nonexistent PCI bus %d!"(%d=[0-2]) messages. I did some grepping to
> find out were this was coming from and find what called it. This is what I
> found:
> 
> pcibios_init()
> pcibios_fixup_irqs()
> IO_APIC_get_PCI_irq_vector()
> 
> I tried commenting pcibios_fixup_irqs() out of pcibios_init() because from
> what I could tell from the kernel logs, IOAPIC had already setup the IRQs.
> This appears to work. Note that the pciirq patch has no effect on this. The
> net effect of the error was that my PCI onboard sound chip and my PCI cards
> had their IRQs set to 0.
> Can anyone find a better solution than my hack? My mobo is an ASUS A7M266-D
> (BIOS v1006) dual Athlon board based on the AMD 760MPX chipset. I can get more
> info if it is needed.
> Thanks in advance.
> 
> --John (who doesn't recall this being discussed before :)