From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:50518)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@redhat.com>) id 1ZoqIf-0004vd-Uc
	for qemu-devel@nongnu.org; Wed, 21 Oct 2015 06:08:04 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <stefanha@redhat.com>) id 1ZoqHY-0007Xl-PD
	for qemu-devel@nongnu.org; Wed, 21 Oct 2015 06:07:17 -0400
Date: Wed, 21 Oct 2015 11:01:24 +0100
From: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20151021100124.GA13408@stefanha-thinkpad.redhat.com>
References: <cover.1443793122.git.berto@igalia.com>
	<87909e13adeb2eb2c71d6613b40d3c9b534db9eb.1443793122.git.berto@igalia.com>
	<20151013153830.GI3057@stefanha-thinkpad.redhat.com>
	<w51bnc01dtg.fsf@maestria.local.igalia.com>
	<20151015145822.GB21733@stefanha-thinkpad.redhat.com>
	<w51zizjm9df.fsf@maestria.local.igalia.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <w51zizjm9df.fsf@maestria.local.igalia.com>
Subject: Re: [Qemu-devel] [PATCH v2 06/22] block: Add "supports_stats" field
	to BlockStats
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: Alberto Garcia <berto@igalia.com>
Cc: Kevin Wolf <kwolf@redhat.com>, qemu-block@nongnu.org, Markus Armbruster <armbru@redhat.com>, qemu-devel@nongnu.org, Max Reitz <mreitz@redhat.com>

On Fri, Oct 16, 2015 at 11:49:00AM +0200, Alberto Garcia wrote:
> On Thu 15 Oct 2015 04:58:22 PM CEST, Stefan Hajnoczi wrote:
> >> > If I/O accounting isn't being used then all fields will be 0?
> >> 
> >> Yes, but there's no way to tell if that happens because I/O
> >> accounting is not supported or because there hasn't been any I/O yet.
> >> 
> >> There's one additional problem: this patch assumes that accounting is
> >> supported if this BDS is attached to a BlockBackend. But we don't
> >> know if the device model supports accounting or not, I still need to
> >> figure out what's the best way to do it.
> >
> > Is there a corresponding libvirt patch or why does it matter whether
> > the QMP client can detect whether blockstats are available?
> 
> I'm thinking that keeping this patch as it is now is probably not very
> useful.
> 
> Block statistics are kept in the BlockBackend, so the only BDS that is
> going to have data != 0 when you call query-blockstats is the topmost
> one. There's probably no need to have an additional flag for this.
> 
> If you disconnect a BlockBackend from a device model that implements
> accounting and then connect it to one that does not, there's no way for
> the client to know that. That's probably worth exposing in the API, but
> this patch does not do that yet, so I think we can skip it for now.

Okay.

Stefan