Remember I mentioned fsresize?  Perhaps it isn't quite ready to be
integrated...but there's _definitely_ a bug in fsck.

I ran fsresize to grow a filesystem.  Upon using the result,

pos = 0, i = 13, fs = /space
panic: ffs_alloccgblk: can't find blk in cyl

So, there's something wrong with fsresize.  That's neither here nor
there...except that I let it come up, and it fscked the filesystem on
the way up...and fsck claims to have fixed it...but it crashed again,
with the same error!

In other words, there's something fsck isn't doing right.  It shouldn't
be possible for a filesystem to pass fsck and then provoke a kernel
panic.

I'm investigating. :-)  So far all I know is what a quick skim of the
ffs_alloccgblk() code tells me, which ain't much.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu