> It seems that there are many problems getting internal video to work on
> the PB1x0 series, and I had an idea last night... if you call
> GetWmgrPort(&g) you get a GrafPtr to the entire screen in g, so would
> g->portBits.baseAddr (I hope I spelled it right) be any help in figuring
> out the video/memory mapping? There is also an element of g->portBits that
> gives the length of the video memory. I can't remember what it's called
> right now.

Well, it's not so much a problem of knowing where the base address of
the screen is (you can find that in the tech notes) or even its size.
The trouble is there is some kind of cooky problem with memory mapping
on powerbooks in general that we're not accounting for.

I think once we've fixed that problem the video will fall in line easily.

If you'd like to see how I fixed the problem in a hacky manner check out
my intvid 0.x patches. They can be found on ftp.macbsd.com in
/pub/outgoing/bjorn/intvidtest/intvidtest.0.x

I haven't looked at them in a while but as I recall, the key to getting
PowerBooks to run was skipping a memory probe of the NuBus regions and
forcing NetBSD to use a serial-boot memory mapping which is apparently
much simpler.