[ From the "ancient history" section of my mailbox ]

> Is there any reason we can't use the "name" field of the partition
> map for our own purposes?

Sort of a "chicken and egg" problem.  Sure, we could have BSD utility
that allowed us to change that partition name, but that won't help
people get NetBSD installed well.  I don't really see a problem using
A/UX partitions if we can figure out what-all they look like.
Actually, BSD doesn't look at the name.  It looks at the partition type,
then at flags within structures of the partition map--namely, the
Block Zero Block.  It occurs to me that we might be too strict here--
looking at the code for the first time in a year or two.  I might play
with this a tad when I find the time.

> It's nice to have the kernel be clever about recognizing partitions,
> but I don't want to have to outsmart it to get it to do what I want. :)
> If we allow people to _tell_ it which sub-device to use for each partition,
> nobody loses.

That was the idea behind allowing the booter to specify a partition
name--that field is ignored unless there are two "root" partitions in
sight on the disk.

-allen

-- 
Allen Briggs - end killing - allen.briggs@bev.net ** MacBSD == NetBSD/mac68k **
   Where does all my time go?  <a href="http://www.netbsd.org/">Guess.</a>