Michael Olbricht wrote;

> I wonder whether it's a just question of kernel config.

Yike.  You're right.  While yet to prepare new config file for it, I
remember that there are two different PMAZ-A drivers in NetBSD
distribution.

	dev/tc/asc_tc.c
	arch/pmax/tc/asc_pmaz.c
		
The former is used by NetBSD/pmax by default.  It's near obsolute
anyway and many DECstation 5000 owners use clean and coherent MI SCSI
version of ASC driver that resides in pmax/tc/ directory.

The old ASC drivers are inherited from 4.4BSD pmax, are to be replaced
by the newer, but it has not been done by a certain reason (Hi, Michael).

So, you need to do a small work to (re-)populate and adjust DEC3000
to use new PMAZ-A driver of pmax/tc/asc_pmaz.c.

> Well, a /300 supplies only two usable drive bays, and I can't make
> the bus too long.

My /300 sports 7 different SCSI devices (6 disks and one 4mm DAT) and
am very pleased with the superior stability.  Disks are striped by
ccd(4) and all of SCSI devices are kicked and ran when dump(8) is
made.  I ripped off internal drives, and hooked everything on outside.
(I use BA350 chasis in this case). 

The reason why PMAD-A can not cope with NetBSD/alpha is the same as
above.  It has little been tested and adjusted for Alpha architecture.
Both are DECstation specific somehow and it's mistake they are populated
in dev/tc/ directory which implies MI.

Tohru Nishimura