Subject: Re: Why don't we trust ACPI for interrupt routing?
To: Frank van der Linden <fvdl@netbsd.org>
From: Martin Husemann <martin@duskware.de>
List: port-i386
Date: 08/26/2006 12:42:40
On Sat, Aug 26, 2006 at 11:21:37AM +0200, Frank van der Linden wrote:
> Try just setting pci_link_bios_isa_irqs at all 1s, or just take out the 
> test for it in acpi_pci_link_choose_irq().  The code seems to prefer 
> IRQs that have been initialized by the BIOS (e.g. ones that have been 
> written into the PCI interrupt line config register).

Thanks for the explanation, Frank. I tried that, but as Simon reported, this
does only partly work. My machine works fine currently with just the fallback
to ACPI SCI commented out (i.e. all my devices sharing IRQ 11) - I'll try to
play some more with it later, but could use any pointers what to try.

Since windows does not seem to special case this (and it is a stock windows
install, nothing Dell specific), there must be a tiny detail we are missing.

If no real fix is found, should we offer a compile time option to disable
the SCI fallback?

Martin