From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([140.186.70.92]:42819)
	by lists.gnu.org with esmtp (Exim 4.71) (envelope-from <hch@lst.de>)
	id 1Qm8uC-0005hH-5S
	for qemu-devel@nongnu.org; Wed, 27 Jul 2011 14:32:29 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <hch@lst.de>) id 1Qm8uB-0000MZ-Bm
	for qemu-devel@nongnu.org; Wed, 27 Jul 2011 14:32:28 -0400
Received: from verein.lst.de ([213.95.11.211]:58572 helo=newverein.lst.de)
	by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from <hch@lst.de>)
	id 1Qm8uB-0000MS-79
	for qemu-devel@nongnu.org; Wed, 27 Jul 2011 14:32:27 -0400
Date: Wed, 27 Jul 2011 20:32:26 +0200
From: Christoph Hellwig <hch@lst.de>
Message-ID: <20110727183226.GB14736@lst.de>
References: <1311791126-11383-1-git-send-email-freddy77@gmail.com>
	<1311791126-11383-3-git-send-email-freddy77@gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1311791126-11383-3-git-send-email-freddy77@gmail.com>
Subject: Re: [Qemu-devel] [PATCH 2/2] aio: use Linux AIO even if nocache
	is	not specified
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: Frediano Ziglio <freddy77@gmail.com>
Cc: Kevin Wolf <kwolf@redhat.com>, qemu-devel@nongnu.org

On Wed, Jul 27, 2011 at 08:25:26PM +0200, Frediano Ziglio wrote:
> Currently Linux AIO are used only if nocache is specified.
> Linux AIO works in all cases. The only problem is that currently Linux AIO
> does not align data so I add a test that use POSIX AIO in this case.

The kernel will accept buffered I/O requests, and even handle them 100%
correctly.  The only thing it won't do is to handle them asynchronously,
so with your patch you're back to executing I/O synchronously in guest
context.