The warning on the NFS kernel patches for NFS client and server locking
has been removed, thanks to some dilligent tracing on the part of
Chris Csanady.

With these patches, modifications to the rpc.lockd, addition of an
F_NONPOSIX fcntl to get rid of the clear-on-close behavior, and
use of F_CNVT patches appropriate for whatever the wire format is
to interface to VOP_FHTOVP, NFS server locking should be completely
functional.

NFS client locking requires that the non-veto nfs_advlock() function
attempt to proxy the lock via RPC to the remote system, and if it fails,
should veto the request.

Note: I have not done anything to deal with blocking NFS lock requests;
IMO, that code belongs in the rpc.lockd.  The F_RSETLKW is specifically
for support of a threaded implementation for proxied blocking locks.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.