Hi,

>>Thanks for the quick reply!  I have compiled a new kernel (just uniprocessor
>>for
>>the moment) with the suggested options and attached the dmesg output.  It looks
>>like the kernel isn't able to figure out the PCI-X and PCI-64 buses on the AMD
>>8131 at all ("failed to derive bus number").  IIRC there should be a second
>>root PCI bus there as well.  It also seems weird that all the interrupts are
>>mapping to isa0.  Is my motherboard firmware simply broken?  Is there anything
>>I can do if it is?
>>
>>BTW, if I enable MPACPI in the kernel options, should I also turn off MPBIOS? 
>>Similarly, are MPACPI_SCANPCI and MPBIOS_SCANPCI mutually exclusive?
> 
> 
> Hmm.. obviously it does need some extra tweaking somewhere (since Windows
> does get it right), but I'm not sure where that is. It seems that the
> BIOS possibly doesn't initialize some PCI bridge registers. Also,
> MP*_SCANPCI are designed to find extra root buses, and those options
> do normally work.

Hmm, I suspect he is a victim of the bug I fixed in November last year in
sys/arch/x86/x86/mpacpi.c, where the bus number on the wrong node is tested.

Unfortunately, I missed to request a pullup of this change to the netbsd-2
branch :-(.  And I'll not be able to do this in a short timeframe either,
as this seems to require a patch.  Maybe, you can?

Otherwise, maybe Chris can test with a (more) current kernel?

Ciao,
Wolfgang
-- 
ws@TooLs.DE                            Wolfgang Solfrank, TooLs GmbH