Excerpts from netbsd: 7-Dec-97 Re: PCI mode probe [was Re:.. Jarkko
Torppa@cute.fi (706*)

> My compaq prosignia says,
> 
> NetBSD 1.3_BETA (COMPA) #1: Sun Nov 30 
> ....
> pchb0 at pci0 dev 0 function 0
> pchb0: vendor 0x0e11 product 0x1000 (rev. 0x01)

And it is found only once? (Not the multiple-function
bug Manuel's "vendor 0x0e11 product 0x4000" has?)

Manuel, can you print out the contents of the
"header type" register ("bhlcr" in pciattach() -
dev/pci/pci.c)? If it contains complete garbage
we can sort out this annoyance cleanly.
And, does your host bridge show up as device
number 0 too?
If yes, and no other "mode-1" machines with don't
pass the strong check show up, we can perhaps
replace the code inside "PCI_PROBE_DONTKNOW"
by a simple Compaq-specific check.

Can you Compaq-owners print out read back
value in pci_machdep.c:pci_mode_detect()
where the first mode-1 check fails?
(perhaps add to the line
"printf("pci_mode_detect: mode 1 enable failed\n");"
inside #ifdef DEBUG)

Thanks & best regards
Matthias