>> is the decstation 5000/200 amongst the listed of boxen to which the pmax
>> port has been ported or will work on ?

Netbsd/pmax works on the 5k/200.  a 200 was the first machine I got
NetBSD bootstrapped on. The two machines I hack on most are a 5k/200
(3MAX) and a 5k/240 (3MAXPLUS).  The 5k/200 has no IO ASIC chip and so
has the cleanest I/O architecture, so it's easy to get right.
However...

>Well, I've never tried to put NetBSD up on a 5k/200 yet, but that model
>is supposed to be supported, probably more so than my 5k/133  ;)
>Since I'm having troubles booting m 5k/133, I might try to make a system
>disk on the 5k/200 and get a kernel on it that will boot both the 200
>and 133. I wonder how tough this will be :|

The 5k/133 (3MIN) has an IO ASIC and scc serial chips, like the
5k/240.  The problems with the 133 are twofold: masking out interrupts
doesn't work on the 3MIN, and the cfb card has a vertical-retrace
interrupt that cannot be disabled on the card; and a mess of problems
with moving the pmax kernel to using rcons and the "standard" console
device-driver, which broke polled console input.  A first install of a
generic kernel requires polled console input.

The other problem I've seen reported is that serial consoles don't
work.  I don't think they ever worked.  The kernel starts off using
PROM functions for console I/O, and doesn't do any clean hand-off
to the kernel driver when the kernel's serial driver  initializes the
serial chips.  I don't have a clean fix for that.

The changes I made yesterday revert to using the old polled console
code for input, and using the "rcons" glass-tty driver for output.  I
have tested that, thoroughly, on a 5k/240, and it should "just work"
on a 5k/133 with a cfb.  If anyone has problems installing today's
miniroot on a 5k/200, a 3100, or a 5k/2x (Personal Decstation), please
let me know.

If someone would volunteer to write up a guide on installing
NetBSD/pmax and put it on the web, that'd be great...

--Jonathan