> On Jan 10,  3:33am, der Mouse wrote:
> >This is the Nth time I've heard something like "This would be neat, but
> >since kernfs (or procfs) is optional, ...".
> >
> >Perhaps it's time to admit that they offer significant useful
> >functionality and make them an if-you-don't-use-it-things-will-break
> >part of the system?

and Steven R. Allen - allen@cssg4.cslab.ds.boeing.com writes
> SGI has done that, and it seems that every other week there's a complaint
> on sgi.admin about something broken.  The answer usually is that someone's
> done a 'umount -a' and the solution is to run 'mntproc' to remount the
> /proc filesystem...
> 
> I have nothing against it myself, but if we do go that route, prominent
> note needs to be made so the clue-impaired have a fighting chance.

how about this solution: have code that relies on /proc, and hack umount so
that -a will umount everything but /proc (or "required" fs's), and 
preserve -a's old behaviour with a new -A option.

# umount -a
# umount -A

d
-- 
David Leonard                            Project Officer, DSTC
The University of Queensland             leonard@dstc.edu.au
http://www.dstc.edu.au/WP/staff/david-leonard.html