From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:39843) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VsaTy-0001jt-GX for qemu-devel@nongnu.org; Mon, 16 Dec 2013 10:53:27 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VsaTr-0000gQ-Nh for qemu-devel@nongnu.org; Mon, 16 Dec 2013 10:53:22 -0500 Received: from mx1.redhat.com ([209.132.183.28]:37465) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VsaTr-0000fC-6F for qemu-devel@nongnu.org; Mon, 16 Dec 2013 10:53:15 -0500 Date: Mon, 16 Dec 2013 16:53:03 +0100 From: Igor Mammedov Message-ID: <20131216165303.15ef73ea@nial.usersys.redhat.com> In-Reply-To: <20131216162625.2e4353c7@nial.usersys.redhat.com> References: <1386938688-22433-1-git-send-email-imammedo@redhat.com> <1386938688-22433-5-git-send-email-imammedo@redhat.com> <20131216162625.2e4353c7@nial.usersys.redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH 04/11] qdev: add to BusState "hotplug-handler" link List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Igor Mammedov Cc: Kevin Wolf , Peter Maydell , Peter Crosthwaite , Eduardo Habkost , "Michael S. Tsirkin" , marcel.a@redhat.com, "qemu-devel@nongnu.org Developers" , Blue Swirl , alex.williamson@redhat.com, Gerd Hoffmann , Anthony Liguori , Paolo Bonzini , dkoch@verizon.co, Andreas =?ISO-8859-1?B?RuRyYmVy?= On Mon, 16 Dec 2013 16:26:25 +0100 Igor Mammedov wrote: > On Sat, 14 Dec 2013 17:05:57 +1000 > Peter Crosthwaite wrote: > [...] > > > > > > QTAILQ_INIT(&bus->children); > > > + object_property_add_link(obj, QDEV_HOTPLUG_HANDLER_PROPERTY, > > > + TYPE_HOTPLUG_HANDLER, > > > + (Object **)&bus->hotplug_handler, NULL); > > > > I think failure of this is fatal. When the patches go through, > > probably want to &error_abort. [...] > > But more to the point, bus could be created during hotplug, and > we do not want to abort guest, instead hotplug operation should fail. > In this case following setting of QDEV_HOTPLUG_HANDLER_PROPERTY > will fail if object_property_add_link() failed before. In following patches there is a couple places asking for being replaced with &error_abort, so if your patches get in before respin, I'll gladly use &error_abort there.