From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:50009)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1Z1vM2-0005rs-Lw
	for qemu-devel@nongnu.org; Mon, 08 Jun 2015 07:36:35 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1Z1vLz-0000xg-44
	for qemu-devel@nongnu.org; Mon, 08 Jun 2015 07:36:34 -0400
Received: from mx1.redhat.com ([209.132.183.28]:33879)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <kraxel@redhat.com>) id 1Z1vLy-0000wl-W0
	for qemu-devel@nongnu.org; Mon, 08 Jun 2015 07:36:31 -0400
Message-ID: <1433763388.5046.32.camel@redhat.com>
From: Gerd Hoffmann <kraxel@redhat.com>
Date: Mon, 08 Jun 2015 13:36:28 +0200
In-Reply-To: <20150608093125.GF30104@stefanha-thinkpad.redhat.com>
References: <3a37931fe0e17adc26bdaa23e33f0aa8.squirrel@gator4003.hostgator.com>
	<20150608093125.GF30104@stefanha-thinkpad.redhat.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] segfault in memcmp
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: Stefan Hajnoczi <stefanha@gmail.com>
Cc: perrier vincent <clownix@clownix.net>, qemu-devel@nongnu.org

On Mo, 2015-06-08 at 10:31 +0100, Stefan Hajnoczi wrote:
> On Fri, Jun 05, 2015 at 05:19:53PM -0500, perrier vincent wrote:
> > Using a very old guest (lenny) with spice and vga=cirrus, I have
> > a segfault:

Hmm, doesn't reproduce on master (booting i386 lenny install iso).
Which qemu version is this?

> > FILE:      ui/spice-display.c
> > FUNCTION:  qemu_spice_create_update
> > LINE:      if (memcmp(guest + yoff + xoff,
> >                        mirror + yoff + xoff,
> >                        bw * bpp) == 0)
> > 
> > The address of mirror + yoff + xoff is out of boundaries.
> > 
> > I use the following to avoid the crash:
> > 
> > ...
> >   img_get_stride = pixman_image_get_stride(ssd->mirror);
> >   img_height = pixman_image_get_height(ssd->mirror);
> >   img_max = img_height * img_get_stride;
> > ...
> >   if (yoff > img_max)
> >     {
> >     if (dirty_top[blk] == -1)
> >       dirty_top[blk] = y;
> >     }

I'm wondering how you end up with yoff being ouf of boundaries in the
first place ...

cheers,
  Gerd