From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:39391)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VMGDa-00057n-3W
	for qemu-devel@nongnu.org; Wed, 18 Sep 2013 07:46:56 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VMGDT-0002m8-H5
	for qemu-devel@nongnu.org; Wed, 18 Sep 2013 07:46:50 -0400
Received: from mx1.redhat.com ([209.132.183.28]:49251)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1VMGDT-0002m0-90
	for qemu-devel@nongnu.org; Wed, 18 Sep 2013 07:46:43 -0400
Message-ID: <523992AD.30507@redhat.com>
Date: Wed, 18 Sep 2013 13:46:53 +0200
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1376554447-28638-1-git-send-email-famz@redhat.com>
	<1376554447-28638-2-git-send-email-famz@redhat.com>
	<20130904123504.GB12733@stefanha-thinkpad.redhat.com>
In-Reply-To: <20130904123504.GB12733@stefanha-thinkpad.redhat.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH v3 1/2] block: allow live commit of active
 image
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: Stefan Hajnoczi <stefanha@gmail.com>
Cc: kwolf@redhat.com, jcody@redhat.com, Fam Zheng <famz@redhat.com>, qemu-devel@nongnu.org, stefanha@redhat.com

Il 04/09/2013 14:35, Stefan Hajnoczi ha scritto:
> Changing from 512 KB to 64 KB can affect performance.  8 times as many
> iops may be issued to copy data.
> 
> Also, the image's cluster size should really be taken into account.
> Otherwise additional inefficiency will be suffered when we populate a
> 128 KB cluster with a COMMIT_BUFFER_SECTORS (64 KB) write only to
> overwrite the remaining part in the next loop iteration.
> 
> This can be solved by setting dirty bitmap granularity to cluster size
> or 64 KB minimum *and* finding continuous runs of dirty bits so larger
> I/Os can be performed by the main loop (up to 512 KB in one request).

At this point, you're basically reinventing the algorithms in
block/mirror.c (even more than this patch is already doing :)).

I wonder if commit-to-active should be internally a separate kind of
job, implemented in block/mirror.c instead of block/commit.c.

Paolo