From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:40934)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <tom.leiming@gmail.com>) id 1X0pPQ-0008BZ-3V
	for qemu-devel@nongnu.org; Sat, 28 Jun 2014 05:59:01 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <tom.leiming@gmail.com>) id 1X0pPP-0003BX-2n
	for qemu-devel@nongnu.org; Sat, 28 Jun 2014 05:59:00 -0400
Received: from mail-vc0-x235.google.com ([2607:f8b0:400c:c03::235]:48363)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <tom.leiming@gmail.com>) id 1X0pPO-0003BR-Uw
	for qemu-devel@nongnu.org; Sat, 28 Jun 2014 05:58:59 -0400
Received: by mail-vc0-f181.google.com with SMTP id il7so5823840vcb.40
	for <qemu-devel@nongnu.org>; Sat, 28 Jun 2014 02:58:58 -0700 (PDT)
MIME-Version: 1.0
In-Reply-To: <53ADE769.3060903@redhat.com>
References: <CACVXFVNVwpcEUeBQr0UsW58B2CG3Qghj-2SSfx9GC1rrwbjw0A@mail.gmail.com>
	<20140627120129.GO12061@stefanha-thinkpad.muc.redhat.com>
	<CACVXFVPYd6Hs9k6x4pVGJpEzLgY2DAJbm02pgc-cOwQH6rhMzg@mail.gmail.com>
	<53ADE769.3060903@redhat.com>
Date: Sat, 28 Jun 2014 17:58:58 +0800
Message-ID: <CACVXFVMYQ62tREcPDs0rh834BuKtygYqLJic71cmuTYjFzgCJg@mail.gmail.com>
From: Ming Lei <tom.leiming@gmail.com>
Content-Type: text/plain; charset=UTF-8
Subject: Re: [Qemu-devel] [regression] dataplane: throughout -40% by commit
	580b6b2aa2
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: Paolo Bonzini <pbonzini@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>, Fam Zheng <famz@redhat.com>, qemu-devel@nongnu.org, Stefan Hajnoczi <stefanha@redhat.com>, "Michael S. Tsirkin" <mst@redhat.com>

On Sat, Jun 28, 2014 at 5:51 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
> Il 27/06/2014 20:01, Ming Lei ha scritto:
>
>> I just implemented plug&unplug based batching, and it is working now.
>> But throughout still has no obvious improvement.
>>
>> Looks loading in IOthread is a bit low, so I am wondering if there is
>> block point caused by Qemu QEMU block layer.
>
>
> What does perf say?  Also, you can try using the QEMU trace subsystem and
> see where the latency goes.

Follows some test result against 8589744aaf07b62 of
upstream qemu, and the test is done on my 2core(4thread)
laptop:

1, with my draft batch patches[1](only linux-aio supported now)
- throughput: +16% compared qemu upstream
- average time spent by handle_notify(): 310us
- average time between two handle_notify(): 1591us
(this time reflects latency of handling host_notifier)

2, same tests on 2.0.0 release(use custom Linux AIO)
- average time spent by handle_notify(): 68us
- average time between calling two handle_notify(): 269us
(this time reflects latency of handling host_notifier)

>>From above tests, looks root cause is late handling notify, and
qemu block layer becomes 4times slower than previous custom
linux aio taken by dataplane.

[1], git://kernel.ubuntu.com/ming/qemu.git #master-0626-batch

Thanks,
-- 
Ming Lei