Chris Jones writes:
> Am I correct in saying that rpc.lockd doesn't support real file locking?

Yes.

> Is anybody planning on fixing this?

No one that I know of is working on it.

> How difficult/complicated is the locking protocol?

It is undocumented, to begin with.

On top of that, there was a math proof that was done several years ago 
that showed that the remote locking semantics could not actually work
right (which might explain why sun has released about fifty million
patched versions of lockd over the years).

None the less, a partially working one would probably be highly
appreciated.

> I'm asking because we have a bunch of (gack) SGI's that need to share home
> directories.  We *could* mount those off another SGI, but that would be
> hideously slow.  The problem with using a NetBSD server, though, is that
> we don't get real file locks, which the SGI's are really adamant about.
> We haven't seen any data corruption yet, but I still worry...
> 
> Chris
> 
> -----------------------------------------------------cjones@math.montana.edu
> Chris Jones                                          cjones@honors.montana.ed
u
>            Mad scientist at large                    cjones@nervana.montana.e
du
> "Is this going to be a stand-up programming session, sir, or another bug hunt
?"
>