>>>>> "IS" == Ignatios Souvatzis <is@jocelyn.rhein.de> writes:

IS> I think the problem is that NFS operates on (filesystem-id, inode) - tuples,
IS> so the server can not distinguish requests. 

Most NFS servers I've seen have file handles with three parts:  a length
cookie, an exporting file system (fsid/inode/gen) tuple, and the file's
(fsid/inode/gen) tuple.  So it is possible to keep track of which file
handles should be usable by which clients, even if the access lists
differ within a file system.

I haven't looked inside NetBSD's file handle generation code yet--maybe
it doesn't make this sort of thing possible?

-- 
==John Kohl <jtk@kolvir.arlington.ma.us>, <john_kohl@alum.mit.edu>
Home page: <http://people.ne.mediaone.net/jtk/>
Bicycling and Skiing to keep fit.