> It depends on exactly what you want to do. Version 1.2.1 is fine on an
> SE/30. I think the April snapshot (which has unfortunately disapeared)
> worked fine. You might try the 970601 snapshot on ftp.netbsd.org.
> 
> Unfortuntatly it seems a bug crawled into the snapshots, and SE/30s have
> video problems (which is weird as they'd worked for years). Watch the
> list in the next few days as we figure out what's up.

Scott or somebody may be trying to put a cleaner version of my Intvid 0.x
patches into the kernel and in the process might have unwittingly broken
something.

However, while I can see how this might break regular intvid I don't see
how it breaks SE/30 video because the SE/30 fakes up it's video as a real
NuBus card in a real NuBus slot ($9 I think) unlike many other intvids
which fake up a card in a non-existant slot $0. Maybe Allen twiddled
something by mistake when he was addint interrupt code to grf_mv?

There are a few things that need some fixing in the video department anyway.
One bad thing is that the rowbytes value for internal video is only valid
if the internal video was the "console" at boot. A few assumptions like
that were fixed with the introduction of the intvid tables in machdep.c
but there are a few other oddities that have to be fixed.

I haven't had a chance to upgrade to -current. I'm still working off of a
1.2 kernel. I may upgrade after I get everything working on intvid and
move the whole mess from direct memory poking to the bus space notation
we're using now. I hope somebody figures out what's going on before then
because I'd like at least the video memory to get mapped in right :)