>Trying to mount /dev/rz1g on /foo under NetBSD gives me:
>
>        rz1: WARNING: using ULTRIX partition information
>        rzopen: ENXIO on rz1g unit 1 part 6
>        # partitions 8, size of 0 = -1325400064
>        /dev/rz1g on /foo: Device not configured
>
>This is very strange, since I'm able to mount /dev/rz1a on /mnt just
>fine.

It looks to me like the  code that reads Ultrix disklabels isn't
reading the Ultrix partition correctly.

NetBSD/pmax uses a 4.4bsd-like disklabel. This has a format very
different from the Ultrix partition label (which lives, if I recall
correctly, in the superblock of an Ultrix 4.2BSD FFS filesystem on
the `a' partition of a disk.)

I didn't write the NetBSD code that reads Ultrix disklabels; though I
*did* tweak an earlier incarnation  of it.  At one point it worked
well enough to boot ECOFF NetBSD kernels from an Ultrix root partition,
and mount Ultrix-created filesystems.  I don't know what broke.
However, 0xb1000000 seems a _huge_ size for a partition. Is your disk
really that big?  What does Ultrix chpt say?


>In the mean time, I'm going to write the base.tar.gz to /dev/rz0d, relabel
>rz0, and hope that NetBSD will be able to read the filesystem.  I can't
>think of what else to try, since the miniroot doesn't come with a mount_nfs
>(or even a route, though ifconfig is there.  Seems kinda strange).

Oof.  That's because they aren't in a 4.4BSD miniroot, either -- or at
least, the 4.4bsd build-a-miniroot script in src/sys/arch/pmax/stand
doesn't include them on a miniroot.