From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=46366 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1PVJcy-00052I-3O for qemu-devel@nongnu.org; Wed, 22 Dec 2010 03:00:52 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1PVJcv-0000sI-To for qemu-devel@nongnu.org; Wed, 22 Dec 2010 03:00:51 -0500 Received: from mail.valinux.co.jp ([210.128.90.3]:35343) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1PVJcv-0000rs-Iq for qemu-devel@nongnu.org; Wed, 22 Dec 2010 03:00:49 -0500 Date: Wed, 22 Dec 2010 17:00:45 +0900 From: Isaku Yamahata Subject: Re: [Qemu-devel] Re: [PATCH] pci: disable migration of p2p bridge Message-ID: <20101222080045.GA7603@valinux.co.jp> References: <20101222062716.GC7814@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20101222062716.GC7814@redhat.com> List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Michael S. Tsirkin" Cc: Blue Swirl , Alex Williamson , qemu-devel@nongnu.org On Wed, Dec 22, 2010 at 08:27:17AM +0200, Michael S. Tsirkin wrote: > On Wed, Dec 22, 2010 at 12:13:43PM +0900, Isaku Yamahata wrote: > > Right now pcibus_get_dev_path() isn't migration save because > > bus number/secondary bus number are set by guest OS. > > So it can't be used reliably for qemu internal id. > > > > For 0.14 release, disable p2p bridge migration at the moment. > > Once pcibus_get_dev_path() is fixed, this patch should be reverted. > > It will be addressed for 0.15 release. > > > > Cc: "Michael S. Tsirkin" > > Cc: Alex Williamson > > Cc: Blue Swirl > > Signed-off-by: Isaku Yamahata > > > Hmm, haven't looked into this deeply - can we do this in one place > when the bridge is created? Unfortunately it's not easy. It requires revising register_device_unmigratable(). I have to admit this patch is ugly. This patch is temporal work around and should be reverted eventually. So I think it is better to address the original issue (allowing migration of p2p bridge) instead of addressing register_device_unmigratable(). thanks -- yamahata