tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: BIOS/ACPI interrupt conflict

On Wed, Feb 09, 2011 at 04:47:12PM -0800, Cliff Wright wrote:
> Bios is correct, and ACPI wrong, I have seen this on other
> machines. And as I said in the 2007 email, even if ACPI had
> been the correct one, it still was not going to setup the
> interrupt.

In this area, and with the current code base, it is very difficult to
say who is wrong... Note that in theory the PCI interrupt link devices
may contain different IRQ sets depending on whether PIC or I/O APIC is
used, but I don't know how well the current code handles this.

> It occurred to me that maybe a test for an apic needs to be
> done. In my case where I have no apic, then the BIOS data
> has to be accepted because nothing else sets up the interrupt.

Yes, if only 8259A PICs are used, probably no calls should be even
made to mess with the (ACPI) PCI interrupt link devices.

I am slowly working with an entirely new implementation, so while the
patch looks reasonable enough, I think it might be best to generally
leave the current regression-prone code intact.

- Jukka.

Home | Main Index | Thread Index | Old Index