From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:57095)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <famz@redhat.com>) id 1b0JJE-0006Ve-1t
	for qemu-devel@nongnu.org; Tue, 10 May 2016 21:51:33 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <famz@redhat.com>) id 1b0JJB-00032v-Qf
	for qemu-devel@nongnu.org; Tue, 10 May 2016 21:51:30 -0400
Date: Wed, 11 May 2016 09:51:24 +0800
From: Fam Zheng <famz@redhat.com>
Message-ID: <20160511015124.GD13488@ad.usersys.redhat.com>
References: <1461030177-29144-1-git-send-email-famz@redhat.com>
	<1461030177-29144-3-git-send-email-famz@redhat.com>
	<20160504101242.GC14972@noname.str.redhat.com>
	<20160505003224.GA13202@ad.usersys.redhat.com>
	<20160506074907.GA5093@noname.redhat.com>
	<20160510032302.GC1207@ad.usersys.redhat.com>
	<20160510083331.GB4921@noname.str.redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20160510083331.GB4921@noname.str.redhat.com>
Subject: Re: [Qemu-devel] [PATCH 2/2] block: Inactivate all children
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: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-devel@nongnu.org, Max Reitz <mreitz@redhat.com>, qemu-block@nongnu.org

On Tue, 05/10 10:33, Kevin Wolf wrote:
> 
> Fair enough. My series didn't have a separate callback, but with yours
> that should be working.
> 
> So is the semantics of .bdrv_inactivate() basically "bdrv_flush, and I
> really mean it"?

Yes.

> 
> > > Maybe we need something like an "active reference counter", and we
> > > decrement that for all children and only call their .bdrv_inactivate()
> > > when it arrives at 0.
> > 
> > That should work, but the effect of the counters are local to one invocation of
> > bdrv_inactivate_all(), and is not really necessary if we do as above.
> 
> Agreed.

Working on another version now.

Fam