Subject: Re: Dynamic SCSI ids (was: A possible way of handling...)
To: Michael Graff <explorer@flame.org>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: tech-kern
Date: 05/12/1997 14:21:07
--------



wrstuden@loki.stanford.edu (Bill Studenmund) writes:

>> Another possability of this scheme is to permit having 4 (maybe 3?)
>> partitions per device. If your system uses lots of ccd's, you can then
>> have twice as many hd's hooked up at once. If ccd's can be partitioned,
>> you get the best of both worlds.

>How about on the i386 we add the concept of slices?  It sure would be
>nice to have a FreeBSD partition, BSDI, and NetBSD all on the same
>disk, and be able to share home and source directories.

My NZ$0.05:

I think that would be a great idea. I have a couple of i386 machines
with NetBSD and Linux partitions. Slices would be a much nicer way to
manage such systems.  As it is, mapping the non-BSD BIOS partitions is
burning burning too many of the per-drive partitions allowed by the
BSD disklabel.

>Using a slice is, on the i386, a BIOS partition, which is then divided
>into the standard ones.
>
>/dev/wd0a would reference the first (and probably only) NetBSD partition
>by default in that case.

>From what I know of FreeBSD and Solaris/x86 use of slices, that
doesn't sound like the right mapping.  Slices should be partitionable
with a BSD-style disklabel at the start of each *BSD partition.  We
should also be able to access `foreign' slices, or partitions of
`foreign' slices.

Sounds to me like we're going to want 32 (BSD) partitions per i386
HDA, allowing up to 8 partitions per slice.  Or possibly more.

This would also let us kill the `d=whole drive, c=bsd parition' misfeature.

Maybe NetBSD could just copy whatever FreeBSD does here.  That would
be a big overall consistency win for those (like Darren Reed) who use both.