Date:   Tue, 23 Apr 2019 10:38:48 -0400
From:   "J. Bruce Fields" <bfields@redhat.com>
To:     NeilBrown <neilb@suse.com>
Cc:     Doug Nazar <nazard@nazar.ca>, linux-nfs@vger.kernel.org
Subject: Re: nfsd bug after f875a792abe9 (nfsd: allow nfsv3 readdir request
 to be larger.)
Message-ID: <20190423143848.GA9302@parsley.fieldses.org>
References: <5b98d3d5-4b80-8783-b212-4fd7ab9b1006@nazar.ca>
 <87tvep6066.fsf@notabene.neil.brown.name>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <87tvep6066.fsf@notabene.neil.brown.name>
User-Agent: Mutt/1.11.3 (2019-02-01)
Sender: linux-nfs-owner@vger.kernel.org
Precedence: bulk

On Tue, Apr 23, 2019 at 10:21:53AM +1000, NeilBrown wrote:
> On Sat, Apr 20 2019, Doug Nazar wrote:
> 
> > After recently upgrading the kernel on one of my servers I was getting 
> > the below crashes. Tracked it down to an NFSv3 root client, bisection 
> > points to f875a792abe9. A revert of that on top of 
> > 5.1.0-rc5-00289-gee2725f711fa works fine.
> 
> hi Doug,
>  thanks for the report and for bisecting!
> 
> I think this bug has been fixed, but the fix hasn't landed upstream yet.
> If you
>    git pull git://git.linux-nfs.org/~bfields/linux.git nfs-next
> 
> you should get
> 
> Commit 3c86794ac0e6 ("nfsd/nfsd3_proc_readdir: fix buffer count and page pointers")
> 
> which should fix this.
> 
> 
> Bruce:  I don't see that patch in your "for-5.1" branch, but you said
>    "Queuing up for 5.1."
> Did something go awry?

Sorry, I've got this confusing system where I commit first to a branch
named for-x.y-incoming, then push it to for-x.y after I've tested it.
Except sometimes I forget to push to the for-x.y branch until the last
minute.

The -incoming branches should always be included in nfsd-next.

Anyway, I'll send a pull request now, it's overdue....

--b.