> Maxpartitions being 8 on the i386 has been a complaint for several
> years now.  [...]

We keep going around and around about this with no resolution.

Personally, I went with stealing another bit for the partition number
quite some time ago on sparc and sun3 and I think alpha; I haven't done
it for i386 yet, but that's for lack of wanting to rather than because
it's hard, as these patches show. :-)  (For myself, I don't give a damn
about a flag day in /dev, so I didn't bother splitting the partition
number bits.)

What I mostly want to say is, there's a quick-fix kludge possible right
now.  Take a big partition and turn it into a one-element ccd, and
(sub)partition that.

In principle I can't see any reason this wouldn't work recursively,
though I've had the occasional bad experience trying to layer things
like vnd and ccd more than one layer deep.

Of course, doing things like this does interesting things to the boot
sequence's assumptions.  I had great fun (FSVO "fun") trying to set up
a system that had barely enough local space to hold a kernel and
minimal utilities, and no root-accessible NFS server.  I ended up doing
an NFS mount of a directory which contained a huge file, imposing a vnd
on the file (which meant tweaking the vnd driver so non-root could
vnconfig), and using that as if it were disk, chrooting to it for
everything.  It mostly worked, though it was fairly slow, and tended to
panic relatively often.  It was, um, "interesting" to get everything to
happen in the right sequence at boot.  (Load kernel, fsck (local disk),
ifconfig, route add default, mount (NFS), su nonroot -c vnconfig, fsck
(vnd), mount (vnd), chroot and go.)

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B