Subject: Re: obio attachment and Macs w/ two or more I/O controllers
To: None <port-macppc@NetBSD.org>
From: Tim Kelly <hockey@dialectronics.com>
List: port-macppc
Date: 11/17/2004 09:18:51
On Tue, 16 Nov 2004 01:18:53 -0500 (EST)
Chris Tribo <ctribo@college.dtcc.edu> wrote:

> Apple Computer MAC-IO I/O Controller (Paddington) (undefined subclass
> 0x00) at pci0 dev 16 function 0 not configured

This problem can go back to the GENERIC config file, which has 

obio0   at pci? dev ? function ?

which hardwires one and only one I/O controller. This bandaid masks the
problem of not properly matching the PCI device to the Open Firmware
node in obio.c. Change it to

obio*   at pci? dev ? function ?

and for the vast majority of the Macs out there, nothing will change.
For those Macs with two or more I/O Controllers, lots of new and
interesting stuff will, though.

With the default wiring of obio0, nothing in the media bay under the
second I/O controller will attach. I have confirmation that the patches
I have for properly wiring obio (using code already in autoconf.c) does
work, as long as there is no device in the second mediabay. Then there
is a problem in that devices in that mediabay don't have their own
interrupt property. 

Apparently the chip involved here is called "Paddington." It might be
one chip with two PCI device IDs, or two separate chips, but one half of
it acts like Heathrow and does not have its own interrupt property, the
other half calls itself Gatwick and _does_ have its own interrupt
property (but no device under does).

tim