From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:54402)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <vrozenfe@redhat.com>) id 1XyGlT-00023s-SA
	for qemu-devel@nongnu.org; Tue, 09 Dec 2014 04:07:33 -0500
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <vrozenfe@redhat.com>) id 1XyGlN-0001IO-OE
	for qemu-devel@nongnu.org; Tue, 09 Dec 2014 04:07:27 -0500
Received: from mx1.redhat.com ([209.132.183.28]:46865)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <vrozenfe@redhat.com>) id 1XyGlN-0001II-Gy
	for qemu-devel@nongnu.org; Tue, 09 Dec 2014 04:07:21 -0500
Message-ID: <1418116022.9785.14.camel@oscar>
From: Vadim Rozenfeld <vrozenfe@redhat.com>
Date: Tue, 09 Dec 2014 20:07:02 +1100
In-Reply-To: <5486B0E1.3090006@huawei.com>
References: 1392796048.11930.8.camel@w7-rhel <5451E0B8.8080607@huawei.com>
	<1414658930.27378.24.camel@oscar> <545212DC.2000503@huawei.com>
	<1414667730.24879.15.camel@oscar> <5452EC0B.1090406@huawei.com>
	<5486B0E1.3090006@huawei.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] question about vioscsi queue depth patch
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: "Wangting (Kathy)" <kathy.wangting@huawei.com>
Cc: qemu-devel@nongnu.org

On Tue, 2014-12-09 at 16:20 +0800, Wangting (Kathy) wrote:
> Hi Vadim,
> 
>     I test the vioscsi 0001-queue-depth254.patch for virtio-win-0.1-74 with minor modifications about
> 
> virtio api, because it have not updated to use the linux api. When a VM with four disks in one controller
> 
> and 4kb 32depth sequential read for each disk at the same time, the tool of iometer which simulates
> 
> the IO model become no response. If fit the patch into virtio-win-0.1-94, it works well.
> 
>     Is there some bug with the old virtio api, or is there some reason that the patch can not work normally
> 
> in virtio-win-0.1-74?
There were a lot of bug-fixes / changes introduced since build 74
(November 2013). Some of them did affect SRB completion and queue depth
logic.
So do you see any performance improvement after applying the patch?

Cheers,
Vadim.

> 
> 
> 
>     Best wishes,
> 
>     Ting Wang
> 
>