From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([208.118.235.92]:46213)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1Ui2HZ-0000bZ-9k
	for qemu-devel@nongnu.org; Thu, 30 May 2013 08:48:46 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1Ui2HS-0003YL-Sd
	for qemu-devel@nongnu.org; Thu, 30 May 2013 08:48:41 -0400
Received: from mail-wg0-x22f.google.com ([2a00:1450:400c:c00::22f]:52298)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <stefanha@gmail.com>) id 1Ui2HS-0003YG-MG
	for qemu-devel@nongnu.org; Thu, 30 May 2013 08:48:34 -0400
Received: by mail-wg0-f47.google.com with SMTP id e11so194718wgh.14
	for <qemu-devel@nongnu.org>; Thu, 30 May 2013 05:48:34 -0700 (PDT)
Date: Thu, 30 May 2013 14:48:30 +0200
From: Stefan Hajnoczi <stefanha@gmail.com>
Message-ID: <20130530124830.GA6169@stefanha-thinkpad.redhat.com>
References: <51A72673.7090300@linux.vnet.ibm.com> <51A74210.3000302@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <51A74210.3000302@redhat.com>
Subject: Re: [Qemu-devel] [RFC] Check backing_file chain's loop
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: Eric Blake <eblake@redhat.com>
Cc: "kwolf@redhat.com" <kwolf@redhat.com>, Dongxu Wang <wdongxu@linux.vnet.ibm.com>, qemu-devel@nongnu.org, stefanha@redhat.com

On Thu, May 30, 2013 at 06:12:00AM -0600, Eric Blake wrote:
> On 05/30/2013 04:14 AM, Dongxu Wang wrote:
> > Hi,
> > 
> > Now block layer does not check whether backing_file chain can be a
> > circle, do you think it is necessary to do a check?
> > 
> > For example,  1.qcow2's backing_file is 2.qcow2 and 2.qcow2's
> > backing_file is 1.qcow2, then any IO operation will be a fault.
> > 
> > If necessary, I will post a patch to fix.
> 
> Yes, it's worth patching.
> 
> > 
> > [$] qemu-img create -f qcow2 1.qcow2
> > qemu-img: Image creation needs a size parameter
> > [$] qemu-img create -f qcow2 1.qcow2 8M
> > Formatting '1.qcow2', fmt=qcow2 size=8388608 encryption=off
> > cluster_size=65536 lazy_refcounts=off
> > [$] qemu-img create -f qcow2 2.qcow2 -o backing_file=1.qcow2
> > Formatting '2.qcow2', fmt=qcow2 size=8388608 backing_file='1.qcow2'
> > encryption=off cluster_size=65536 lazy_refcounts=off
> > [$] qemu-img create -f qcow2 1.qcow2 -o backing_file=2.qcow2
> > Formatting '1.qcow2', fmt=qcow2 size=8388608 backing_file='2.qcow2'
> > encryption=off cluster_size=65536 lazy_refcounts=off
> > [$] qemu-io -c "read -v 0 512" 1.qcow2
> > ^CSegmentation fault (core dumped)
> 
> Also make sure that 'qemu-img info 1.qcow2' and 'qemu-img info
> --backing-chain 1.qcow2' don't crash.

qemu-img info --backing-chain detects cycles and returns an error, see
the hash table in collect_image_info_list().

We should protect bdrv_open() too.

Stefan