tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

re: basic support for (software concept) "pci domains" in the MI pci code

   On Mon, Aug 24, 2009 at 05:06:07AM +1000, matthew green wrote:
   > sometimes they aren't parents but peers -- eg, psycho case.
   > there's an easy way to figure out which is bus A and bus B, but
   > they aren't parent/children in the autoconf sense.
   I still don't get what a "pci domain" is. In the psycho case, both pci buses
   are completely independend (at least outside the MD kernel parts that deal
   with the bridge setup), so I would have counted them both as their own
   domain, which matches with their parent not being a pci device.

in the way i see it, a single psycho* chip implements a single
domain -- which has two pci busses.  on multiboard systems, each
board with a psycho* would be a different domain.

separate domains exist when pci bus numbers are re-used.  eg,
when something has two "bus 0"'s.  i understand that happens on
macppc platforms.


Home | Main Index | Thread Index | Old Index