Re: problem with sio_pci_fixup(noname_map_irq, 0x0b0a0f0d) under alpha

Thomas Bogendoerfer (tsbogend@alpha.franken.de)
Thu, 9 Sep 1999 22:26:29 +0200

On Wed, Sep 08, 1999 at 10:28:04PM -0400, Peter Sjöberg wrote:
> I did compare and reaply all fixes between 2.2.0-pre7 and 2.2.0 and manage
> to nail it down to a small patch in alpha/kernel/sys_sio.c to fix it even
> under 2.2.12. In the noname_pci_fixup function the sio_pci_fixup call
> changed from ...0d to ...0e and after revers that change it worked fine
> again. Why was that changed? what does it do?

it maps one of the PCI slots interrupts to 13 instead of 14. When mapped
to 14, the IDE support doesn't work.

> - sio_pci_fixup(noname_map_irq, 0x0b0a0f0d);
> + sio_pci_fixup(noname_map_irq, 0x0b0a0f0e);

this patch will break IDE support. I'm using

sio_pci_fixup(noname_map_irq, 0x0b0a0f09);

which will break sound on the UDB again. Mapping the PCI interrupt to 13
is a bug, because it's not possible with the SIO ISA bridge. I've no
idea, which piece will break, when we map it to 5 or 7. It might work
to share a interrupt (depends on the used devices/drivers), but I haven't
tried that.

Thomas.

-- 
   This device has completely bogus header. Compaq scores again :-|
It's a host bridge, but it should be called ghost bridge instead ;^)
                                        [Martin `MJ' Mares on linux-kernel]


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/