Subject: Re: PCI on Compaq deskpro 4000 ?
To: Jason Thorpe <thorpej@nas.nasa.gov>
From: Matthias Drochner <drochner@zelux6.zel.kfa-juelich.de>
List: port-i386
Date: 10/16/1997 14:09:43
Excerpts from netbsd: 16-Oct-97 Re: PCI on Compaq deskpro 4.. Manuel
BOUYER@antioche.l (669*)

> > Now that the BIOS trampoline has been generalized a bit, pci_mode_detect()
> > should probably be changed to ask the PCI BIOS.
> 
> Would be nice if this can avoid this kind of problem. But this may introduce
> new problems due to broken PCI biosses ...

I'd like to avoid BIOSes as much as possible. PCI BIOSes
which fail to report the configuration mode might be rare, but
in general, hardware improves, but BIOSes stay a relict from
8086 times - misdesigned and broken.

We should probably check for mode 1 first, almost all
machines use this. Perhaps the checks should look at more
details too. (The ultimative check would be to try all
device IDs and take existence if at least 1 device as success.)
Are there people with (probably older) mode 2 PCI PCs
who can test some methods?

best regards
Matthias