Date: Thu, 28 Sep 2017 11:34:44 +0100
From: Stefan Hajnoczi <stefanha@redhat.com>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: NeilBrown <neilb@suse.com>,
        "Daniel P. Berrange" <berrange@redhat.com>,
        Chuck Lever <chuck.lever@oracle.com>,
        Steven Whitehouse <swhiteho@redhat.com>,
        Steve Dickson <SteveD@redhat.com>,
        Linux NFS Mailing List <linux-nfs@vger.kernel.org>,
        Matt Benjamin <mbenjami@redhat.com>,
        Jeff Layton <jlayton@redhat.com>,
        Justin Mitchell <jumitche@redhat.com>
Subject: Re: [PATCH nfs-utils v3 00/14] add NFS over AF_VSOCK support
Message-ID: <20170928103444.GD12157@stefanha-x1.localdomain>
References: <20170921170017.GK32364@stefanha-x1.localdomain>
 <d760ecdc-988d-8145-f816-6faf8f2abf03@redhat.com>
 <E5B4D2FF-0DB4-417D-998A-6BBFC36C35C5@oracle.com>
 <20170922115524.GN12725@redhat.com>
 <87efqu6wl4.fsf@notabene.neil.brown.name>
 <20170926034026.GA19283@fieldses.org>
 <20170926133949.GB25286@fieldses.org>
 <20170926134239.GC25286@fieldses.org>
 <20170927122258.GC14579@stefanha-x1.localdomain>
 <20170927134634.GB9585@fieldses.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <20170927134634.GB9585@fieldses.org>
Sender: linux-nfs-owner@vger.kernel.org

On Wed, Sep 27, 2017 at 09:46:34AM -0400, J. Bruce Fields wrote:
> On Wed, Sep 27, 2017 at 01:22:58PM +0100, Stefan Hajnoczi wrote:
> > On Tue, Sep 26, 2017 at 09:42:39AM -0400, J. Bruce Fields wrote:
> > > By the way, do we know anything about likely performance of NFS/VSOCK?
> > 
> > virtio-vsock is designed for reliable host<->guest communication, not
> > performance.  It is not a fast-path to avoid Ethernet/IP.  I haven't run
> > benchmarks on NFS over AF_VSOCK but don't expect its performance to set
> > it apart from virtio-net.
> 
> OK.
> 
> But if we implement NFS/VSOCK and it turns out to be a success, I expect
> people will start using it for things that weren't expected and
> complaining about performance issues.
> 
> I guess I'm not too concerned about performance of the initial
> implementation but it'd be nice to know that there's the possibility to
> optimize later on.
> 
> But if our answer will be just to go figure out how to use a proper
> NFS/TCP mount instead then I suppose that's OK.

Yes, virtio-vsock can be extended in the future for performance
optimizations.

Stefan