briggs@NetBSD.org wrote:

> On Fri, May 12, 2006 at 05:19:59AM +0900, Izumi Tsutsui wrote:
> > Why should we skip it explicitly?
> 
> In some cases (I forget which), probing the PCI side of the PCI Host
> Bridge from the bridge itself can do bad things like hang the system
> or cause an exception.

Including accesses against PCI configuration space of the bridge itself?
If so, I guess we can handle it simply to have some check against
bus/dev/function in pci_conf_read(9) like cobalt/pci/pci_machdep.c.

But according to shige's post, the bridge on 405GP was probed and
"not configured" currently so a simple match function (which checks
ID_REG or CLASS_REG) should work, and there is few thing to be done
in pchb(4) at pci side anyway so no further probing is needed on it.

BTW, current powerpc/ibm4xx/pci/pchb.c (which configures the bridge
at localbus) use pci_conf_read(9) in the match (probe) function,
but it seems ugly a bit for me. Isn't it enough to check CPU ID
(or something), as Jason said?
---
Izumi Tsutsui