From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:49147)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <flukshun@gmail.com>) id 1UIQQf-0007zh-Lv
	for qemu-devel@nongnu.org; Wed, 20 Mar 2013 17:20:14 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <flukshun@gmail.com>) id 1UIQQe-00078z-HU
	for qemu-devel@nongnu.org; Wed, 20 Mar 2013 17:20:13 -0400
Sender: fluxion <flukshun@gmail.com>
Date: Wed, 20 Mar 2013 16:15:59 -0500
From: mdroth <mdroth@linux.vnet.ibm.com>
Message-ID: <20130320211559.GF1580@vm>
References: <1362159627-20139-1-git-send-email-mdroth@linux.vnet.ibm.com>
	<20130320105451.0df45933@doriath> <20130320160308.GB1580@vm>
	<20130320125851.2e501c00@doriath> <20130320172630.GC1580@vm>
	<20130320134056.464e09b6@doriath> <20130320181421.GD1580@vm>
	<20130320143835.4a28e14d@doriath> <20130320193955.GE1580@vm>
	<20130320155618.336f1171@doriath>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20130320155618.336f1171@doriath>
Subject: Re: [Qemu-devel] [PATCH] qemu-ga: use key-value store to avoid
 recycling fd handles after restart
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: Luiz Capitulino <lcapitulino@redhat.com>
Cc: qemu-devel@nongnu.org, aliguori@us.ibm.com, qemu-stable@nongnu.org, armbru@redhat.com

On Wed, Mar 20, 2013 at 03:56:18PM -0400, Luiz Capitulino wrote:
> On Wed, 20 Mar 2013 14:39:55 -0500
> mdroth <mdroth@linux.vnet.ibm.com> wrote:
> 
> > So I wonder if, rather than pursuing option 3, we just introduce an
> > interface that does what we really want and returns handles as UUIDs,
> > then mark the existing interfaces as deprecated (and then remove them
> > within the next 300 years so our assert never gets hit :)
> 
> This seems reasonable to me. It should be an assert() if it should never
> happen. If we feel like doing option 3, we introduce a new interface
> that handles UUIDs instead.
> 

Ok, I think it's a plan then. I have a partial implementation for
guest-file-* commands on windows, and a posix/w32 guest-exec-async
implementation that re-uses them, so maybe those should be introduced
around the new interfaces to avoid churn. Need to generate uuid handles
for guest-exec handles anyway.