Replying to my own message...

Yesterday I compiled a kernel from 1.3.3 fresh sources (GENERIC and
SUPERHYPERTURBO-OPTIMIZED) and to my amazement they both worked with my
dcm-card. This was probably because I changed the ipl from 5 to 3 and select
code from 13 to 16. After kicking myself in the head for changing two (or
three if you count the kernel) variables at the same time, I changed the ipl
back to 5. Still worked, so it was probably the select code.

Now I have *no* idea why it didn't work with sel.code 13 since it was
definitely available:
topcat0 at default and topcat1 at 193, internal hp-ib at default, internal
rs-232 at 9, SCSI S 32 at 14, LAN at 21, PARALLEL at 23 and SCSI S 16 at 30.
(uh yes, the machine is a 380)
There's probably some kind of a rule here, but I have no idea what...

Anyway, it works now and that's the main thing!

*misc.rambling+advocacy follows:*
My brother just installed Redhat 5.something and that led me to think "why
do
I like netbsd ?". Well, I was about to find out and fast.. It seems the
default RH installation comes with kazillion different things and no mere
mortal can never grasp all the stuff that's there. At least I can't.

I remember when eons ago I installed SCO Xenix 2.something on my 286.
The partition was about 20MB and there was plenty of space left on the drive
after a complete installation. The main point is that I *knew* what was
there. I
knew the location of every single snippet, which was nice.

I get the same warm and fuzzy feeling with NetBSD, as the default
installation is comfortably small and if I want X, I know where to find it.
Try that with <insert some linux-distribution here>,
And there's also the fact that I prefer the BSD-way of things...
So to make the story short, I like this OS, I like port-hp300 and keep up
the good work!

--
Jarkko Teppo
jarkko.teppo@nfp.fi