>>> [/dev/sdXX symlink to something like /dev/scsi/sXcStX]
>> In fact, I'd go further and call it really gross.  One of the things
>> I hate most about our SGI machines is these long baroque disk names.
> Well, one of the things I hate about those BSD systems is their flat
> /dev directory.  No structure whatsoever (excpet for the /dev/fd
> subdirectory).  I'd love to have /dev/scsi, /dev/pty, /dev/ipi and so
> on.

How people differ.  I _like_ a flat /dev, though I admit I'm not sure
why.

You can, y'know, lock down sd* to scsi device assignments in your
kernel, and then put the /dev nodes in /dev/scsi/whatever instead of
/dev/sd*.  AFAICT this will give you almost all of what you want,
except having to know about the sdNNN names when mucking about in your
kernel.

> Given the time it can take to download the source of the kernel, it's
> a real annoyance.  Until now, the only reason why I needed that
> source was to add /dev/sd1b as a swap device.

Eek.  I'm astonished anyone would even consider not having kernel
source around (given enough disk space to build kernels, that is).
(Not that that excuses the requirement that every possible swap device
be compiled into the kernel, of course.)

>>> or the (supposed) inability of the normal kernel to mount anything
>>> else than /dev/sd0a on /.
>> I think "config netbsd swap generic" (you actually want "root
>> generic", but they go together - something else that needs fixing)
>> works now, though I'm not sure.
> But supposedly this generic option prevents the use of a second swap
> device.

Probably; I rather doubt "swap on generic and sd1b" would work.

A good project for some ambitious kernel hacker. :-)

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu