> I agree. This issue crossed my mind when I integrated the UFS2 code,
> but I forgot about it. Thanks for picking it up.

I'm not sure, but I might even have suffered some of these problems!

Another interesting question.
If I create an FFSv1 filesystem with less than 1792 (or so) inodes
per cylinder group, then write a large file that consists solely of
the value of the FFSv1 or FFSv2 magic number (or maybe even a validly
constructed superblock) - what will happen next time someone tries to
mount the filesystem?

	David

-- 
David Laight: david@l8s.co.uk