ron minnich wrote:
On Fri, Aug 29, 2008 at 9:27 AM, Stefan Reinauer stepan@coresystems.de wrote:
a) a pci domain
that's my error. Basically, I figured at the time that
- everything was going PCIe
- it was simpler to just have a domain on everything, and have it be
"empty" and 0 on non-PCIe systems for what little time we had them. I.e. rather than have #ifdef everywhere just have a fixed set of rules as to the hierarchy.
b) a pci bus "device" whatever that is supposed to depict c) a pci bridge (host bridge)
In many cases, we have chips that have a device function and a bridge function. Again, I may have made a mistake when I moved the code over. But there really are chips that have both functions.
hm.. so do we need the "bus" construct at all? As far as I understand you, a bus hangs off of either a bridge or a pci domain (which is kind of a top level bridge between the FSB and PCI)?
Just wondering, the more stuff we can toss out, the easier it might/will get.