From mboxrd@z Thu Jan  1 00:00:00 1970
From: Shirley Ma <mashirle@us.ibm.com>
Subject: Re: vhost-net patches
Date: Wed, 28 Oct 2009 09:45:37 -0700
Message-ID: <1256748337.6433.11.camel@localhost.localdomain>
References: <20091023110438.GA20229@redhat.com>
	 <1256310168.4443.2.camel@localhost.localdomain>
	 <1256310765.4443.4.camel@localhost.localdomain>
	 <1256315020.4443.12.camel@localhost.localdomain>
	 <20091026200513.GA26623@redhat.com>
	 <1256592889.10142.8.camel@localhost.localdomain>
	 <20091027064302.GB26914@redhat.com>
	 <1256654819.4753.6.camel@localhost.localdomain>
	 <20091027152753.GA4622@redhat.com>
	 <1256661378.6745.2.camel@localhost.localdomain>
	 <20091028153859.GA28926@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Cc: Sridhar Samudrala <sri@us.ibm.com>, Shirley Ma <xma@us.ibm.com>,
	David Stevens <dlstevens@us.ibm.com>, kvm@vger.kernel.org,
	sri@linux.vnet.ibm.com, mashirle@linux.vnet.ibm.com
To: "Michael S. Tsirkin" <mst@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from e32.co.us.ibm.com ([32.97.110.150]:35477 "EHLO
	e32.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1755114AbZJ1QqC (ORCPT <rfc822;kvm@vger.kernel.org>);
	Wed, 28 Oct 2009 12:46:02 -0400
Received: from d03relay02.boulder.ibm.com (d03relay02.boulder.ibm.com [9.17.195.227])
	by e32.co.us.ibm.com (8.14.3/8.13.1) with ESMTP id n9SGevJ6014777
	for <kvm@vger.kernel.org>; Wed, 28 Oct 2009 10:40:57 -0600
Received: from d03av01.boulder.ibm.com (d03av01.boulder.ibm.com [9.17.195.167])
	by d03relay02.boulder.ibm.com (8.13.8/8.13.8/NCO v9.1) with ESMTP id n9SGjnKu126462
	for <kvm@vger.kernel.org>; Wed, 28 Oct 2009 10:45:51 -0600
Received: from d03av01.boulder.ibm.com (loopback [127.0.0.1])
	by d03av01.boulder.ibm.com (8.14.3/8.13.1/NCO v10.0 AVout) with ESMTP id n9SGjdY6003291
	for <kvm@vger.kernel.org>; Wed, 28 Oct 2009 10:45:39 -0600
In-Reply-To: <20091028153859.GA28926@redhat.com>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

Hello Michael,

On Wed, 2009-10-28 at 17:39 +0200, Michael S. Tsirkin wrote:
> Here's another hack to try. It will break raw sockets,
> but just as a test:

This patch looks better than previous one for guest to host TCP_STREAM
performance. The transmission queue full still exists, but TCP_STREAM
results is 43xxMb/s (userspace is about 3500Mb/s). When I increase the
transmission queue size to 1K, The performance burst up to 53xxMb/s. But
some runs I still hit several transmission queue full in 60 secs run
even with 1K queue size.

Thanks
Shirley