From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([140.186.70.92]:48478) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Rxf2d-00069t-Ao for qemu-devel@nongnu.org; Wed, 15 Feb 2012 08:37:04 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Rxf2Z-0002x8-3d for qemu-devel@nongnu.org; Wed, 15 Feb 2012 08:37:03 -0500 Received: from mx1.redhat.com ([209.132.183.28]:12376) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Rxf2Y-0002wp-TH for qemu-devel@nongnu.org; Wed, 15 Feb 2012 08:36:59 -0500 Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id q1FDaw7f009828 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Wed, 15 Feb 2012 08:36:58 -0500 Message-ID: <4F3BB4F7.3030906@redhat.com> Date: Wed, 15 Feb 2012 14:36:55 +0100 From: Gerd Hoffmann MIME-Version: 1.0 References: <1329311466-20344-1-git-send-email-kraxel@redhat.com> <4F3BB1A0.7050909@redhat.com> In-Reply-To: <4F3BB1A0.7050909@redhat.com> Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH] qxl: don't render stuff when the vm is stopped. List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Yonit Halperin Cc: qemu-devel@nongnu.org On 02/15/12 14:22, Yonit Halperin wrote: > Hi, > On 02/15/2012 03:11 PM, Gerd Hoffmann wrote: >> This patch fixes the local qxl renderer to not kick spice-server in case >> the vm is stopped. First it is pointless because we render evevything >> when the vm is stopped. Thus there is nothing to render anyway because >> a stopped guest can hardly queue more commands. > hmm...When the vm is stopped we render only commands that we already > read. We don't do more reading from the command ring. Ah, ok. > So there may be > other pending commands that were not rendered. That is why I > suggested allowing rendering when the vm is stopped. But > not allowing it during loading the vm during migration. I'd prefer to keep things simple: let the spice worker run when the guest runs, no exceptions. We may leave some unrendered commands in the ring then. Ok. Is that a problem for some reason? Note that the guest may have more commands queued which spice-server simply doesn't see yet due to the ring being full. If we stop the guest the wrong moment we can end up with a half-done screen update operation no matter what. cheers, Gerd