Re: [PATCH 00/19] OPEN optimisations and Attribute delegations

[Trond Myklebust <trondmy@hammerspace.com>]

On Fri, 2024-06-14 at 23:27 -0700, Christoph Hellwig wrote:
> On Fri, Jun 14, 2024 at 08:34:32AM -0400, Jeff Layton wrote:
> > This all looks pretty reasonable except for the last two patches.
> > Probably, they should be squashed together since there is no caller
> > of
> > ->return_delegation until the last one. There is also nothing
> > describing the changes there, and I think it could use some
> > explanation
> > (though I think I get what you're doing).
> > 
> > Finally, I suppose we need to look at implementing support delstid
> > in
> > knfsd as well. I'll open a new feature request for that the linux-
> > nfs
> > project on github.
> 
> I don't think there ever was a formal rule, but having a feature like
> this that affects all of the core nfs code without beeing able to
> test
> it against knfsd seems a bit dangerous indeed.
> 

We have been going through the IETF process to ensure there is an
agreed upon protocol for this that can be used to test the client code
and validate its correctness. That spec can also be used by others to
contribute the missing knfsd code.
At some point soon, I do hope that Hammerspace will be in the position
where we're able to contribute both client and server code for these
features; for now, we are unfortunately still resource constrained.

-- 
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com