> I think the introduction of the obio approach was very close to the
> time that the failure to probe started to appear.

The introduction of obio being somewhat responsible is my guess as
well.  Following Allen's advise, I turned DEBUG macro on in nubus.c
for clues.  Guess what happened?  My two video devices were probed and
atached as grf1 and 2 respectively.  No grf0 which is described to be
"at obio?" in my config file as well as in GENERIC.  I'm trying
another kernel from the same config except the commented-out grf0
line.

ken