From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:47317)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1VnrNm-0002OV-H1
	for qemu-devel@nongnu.org; Tue, 03 Dec 2013 09:55:34 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1VnrNd-00023W-M0
	for qemu-devel@nongnu.org; Tue, 03 Dec 2013 09:55:26 -0500
Received: from mail-wi0-x229.google.com ([2a00:1450:400c:c05::229]:42037)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1VnrNd-00023B-GA
	for qemu-devel@nongnu.org; Tue, 03 Dec 2013 09:55:17 -0500
Received: by mail-wi0-f169.google.com with SMTP id hn6so2142790wib.4
	for <qemu-devel@nongnu.org>; Tue, 03 Dec 2013 06:55:16 -0800 (PST)
Date: Tue, 3 Dec 2013 15:55:12 +0100
From: Stefan Hajnoczi <stefanha@gmail.com>
Message-ID: <20131203145512.GH24604@stefanha-thinkpad.muc.redhat.com>
References: <CAJSP0QXLfmq_dPvh90s-5VfaMN48u66+teca5_qYCq+ROpO=tw@mail.gmail.com>
	<20131128021513.GJ2360@kernel.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20131128021513.GJ2360@kernel.dk>
Subject: Re: [Qemu-devel] Linux multiqueue block layer thoughts
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Jens Axboe <axboe@kernel.dk>
Cc: Kevin Wolf <kwolf@redhat.com>, Paolo Bonzini <pbonzini@redhat.com>, qemu-devel <qemu-devel@nongnu.org>, Michael Roth <mdroth@linux.vnet.ibm.com>

On Wed, Nov 27, 2013 at 07:15:13PM -0700, Jens Axboe wrote:
> On Wed, Nov 27 2013, Stefan Hajnoczi wrote:
> > At the end of all this we'd arrive at the following architecture:
> > 1. Guest virtio device has multiple queues (1 per node or vcpu).
> > 2. QEMU has multiple dataplane/QContext threads that process virtqueue
> > kicks, they are bound to host CPUs/nodes.
> > 3. Linux kernel has multiqueue block I/O.
> 
> I think that sounds very reasonable. Let me know if there's anything you
> need help or advice with.
> 
> > Jens: when experimenting with multiqueue virtio-blk, how far did you
> > modify QEMU to eliminate global request processing state from block.c?
> 
> I did very little scaling testing on virtio-blk, it was more a demo case
> for conversion than anything else. So probably not of much use to what
> you are looking for...

Okay, thanks.  It will be a while before the whole stack supports
multiqueue but it's good to know this approach sounds reasonable.

Stefan