From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from [140.186.70.92] (port=55688 helo=eggs.gnu.org)
	by lists.gnu.org with esmtp (Exim 4.43) id 1P5FSk-00058E-Sv
	for qemu-devel@nongnu.org; Mon, 11 Oct 2010 06:18:35 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <kwolf@redhat.com>) id 1P5FSk-00043k-02
	for qemu-devel@nongnu.org; Mon, 11 Oct 2010 06:18:34 -0400
Received: from mx1.redhat.com ([209.132.183.28]:58843)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <kwolf@redhat.com>) id 1P5FSj-00043d-QL
	for qemu-devel@nongnu.org; Mon, 11 Oct 2010 06:18:33 -0400
Message-ID: <4CB2E498.3070008@redhat.com>
Date: Mon, 11 Oct 2010 12:19:04 +0200
From: Kevin Wolf <kwolf@redhat.com>
MIME-Version: 1.0
References: <m3lj65122k.fsf@blackfin.pond.sub.org>
In-Reply-To: <m3lj65122k.fsf@blackfin.pond.sub.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Subject: [Qemu-devel] Re: virtio_cleanup() missing in virtio-{blk, balloon,
	9p}-pci
List-Id: qemu-devel.nongnu.org
List-Unsubscribe: <http://lists.nongnu.org/mailman/listinfo/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: <http://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Markus Armbruster <armbru@redhat.com>
Cc: Venkateswararao Jujjuri <jvrao@linux.vnet.ibm.com>, qemu-devel@nongnu.org, Adam Litke <agl@us.ibm.com>

Am 11.10.2010 12:00, schrieb Markus Armbruster:
> Pattern:
> 
> FOO's PCIDeviceInfo method exit is virtio_FOO_exit_pci(), which calls
> virtio_FOO_exit() and virtio_exit_pci().
> 
> virtio_{net,serial}_exit() call virtio_cleanup().
> 
> virtio_blk_exit() doesn't.  Why?
> 
> virtio-balloon-pci uses virtio_exit_pci() as exit method.  No
> virtio_cleanup()?
> 
> virtio-9p-pci doesn't have an exit method.  Doesn't feel right.
> 
> I suspect these three leak memory or worse on hot unplug.

Yes, virtio-blk seems to leak memory there.

We should probably also cancel all running requests before removing the
disk. I suspect not doing this could cause some segfaults when the AIO
callback wants to access a disk that doesn't exist any more.

Kevin