In message <199910181736.KAA08025@Greyhawk.DSG.Stanford.EDU>,
Jonathan Stone writes:

>In message <199910180528.BAA02244@ginger.cmf.nrl.navy.mil> 
>Ken Hornstein writes

>>Thanks for the information; I had feared this was going to happen
>>eventually (a PC which doesn't do the right thing when "Plug and
>>Play OS" was set to no).  I was hoping in Jonathan's case it would
>>get him going.
>
>Ken,

[...]

>What I want is to get the cardbus code going.  For that, both PnPoS=no
>and the PCIBIOS patch set all the previousuly- unconfigured-IRQ PCI
>devices to IRQ 9. Which may be correct for the Ricoh PCi-Cardbus
>bridge on this Vaio; I can' tell, yet.

Okay.  Back at work, with a known-good 3c589, *and* with the `pcic0 at
isa? port 0x320' commented out,' the CardBus config, plus PCIBIOS
patches, works regardless of whether the Phoenix NoteBios `PnP OS' is
set to `yes' or `no'. And with and without Soren's patch.

I'm guessing my problems were with the pcic0 legacy-comatiblity ports
being touched, but it's just a guess.  Configuring pcic0 does seem to
clobber the PCI IRQ on the Cardbus bridge, even if the `non-PnP' BIOs
setting set it up.  Seems consistent, anyway.

So when can we get those i386 PCIBIOS patches into the tree as the
approved MD-pci-setup for i386?