> A long time ago I also discussed a problem where the . and .. entries
> of a directory would seem to be the same inode, and other obvious
> wrong inodes were present.

This reminds me of a run-in I had with some commercial NFS-server box.
I think it was a Network Appliance and the problem was related to their
snapshot directories, but can't recall details (and I've long since
left the environment where it occurred).  I think it was handing out
non-unique file IDs within a single mount (though with different
filehandles).  Confused userland programs that expected <st_dev,st_ino>
pairs to be unique.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B