Ok, Joe Laffey and I just about have the second I/O controller stuff
worked out. I've got the two I/O controllers attaching properly, and the
mediabay under the second I/O controller is close to attaching properly.
I hope to have a final patch within a day or two, along with the patch
to fix the dropping of nodes in the boot path identification.

However, there is a problem with the second I/O controller, as it is
responsible for the interrupt. This causes problems when there is a hard
drive in the mediabay, because the hard drive node matches wdc and then
one gets

mediabay0 at obio0 offset 0x34
wdc0 at mediabay0 offset 0x21000: couldn't get irq property relying on
parent irq 27

(there is a bit of editing, this is from the testing)

I can not see where wdc is allowed to attach to a mediabay in the
files.macppc file. It appears to me that because both mediabay and wdc
matching routines are present, the config_found routine calls both.

So my question is, is there some mechanism that ensures that device
attachment is consistent with the config file of the kernel?

thanks in advance,
tim