Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
MIME-Version: 1.0
From:   "NeilBrown" <neilb@suse.de>
To:     "Chuck Lever III" <chuck.lever@oracle.com>
Cc:     "Jeff Layton" <jlayton@kernel.org>,
        "Linux NFS Mailing List" <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH 00/13] NFSD: clean up locking
In-reply-to: <44AD17F0-FC0B-4F20-BCED-071CD70DFBBC@oracle.com>
References: <165881740958.21666.5904057696047278505.stgit@noble.brown>,
 <44AD17F0-FC0B-4F20-BCED-071CD70DFBBC@oracle.com>
Date:   Fri, 29 Jul 2022 11:29:11 +1000
Message-id: <165905815102.4359.6022541493569857679@noble.neil.brown.name>
Precedence: bulk

On Fri, 29 Jul 2022, Chuck Lever III wrote:
> 
> > On Jul 26, 2022, at 2:45 AM, NeilBrown <neilb@suse.de> wrote:
> > 
> > This is the latest version of my series to clean up locking -
> > particularly of directories - in preparation for proposed patches which
> > change how directory locking works across the VFS.
> > 
> > I've included Jeff's patches to validate the dentry after getting a
> > delegation.  The second patch has been changed quite a bit to use
> > nfsd_lookup_dentry(). I've left Jeff's From: line in place - let me know
> > if you'd rather I change it.
> > 
> > Setting of ACLs and security labels has been moved from nfs4 code to
> > nfsd_setattr() which allows quite a lot of code cleanup.
> > 
> > I think I've addressed all the concerns that have been raised, though
> > maybe not in the way that was suggested.
> > 
> > I've tested this with cthon tests over v2, v3, v4.0, v4.1, and xfstests
> > on v3 and v4.1, and pynfs 4.0, 4.1.  No problems appeared.
> > 
> > Thanks,
> > NeilBrown
> 
> Hi Neil-
> 
> No objections to this round, looks like a very good set of clean-ups.
> 
> I've also resurrected NFSv2 on my test systems, and captured a baseline
> without these patches applied.
> 
> There are a number of cosmetic issues I found, posting those separately.
> If you trust me, I can take care of those here, or you can submit a
> v3 (v4?).

I would love for you to make those changes yourself!
As I noted separately there is a bug : nfserrno() needed where you
suggested __be32.
All others are cosmetic and I trust you to fix those up however seems
best.

Thanks,
NeilBrown