From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Graf Subject: Re: [PULL] virtio Date: Sun, 7 Mar 2010 12:11:26 +0100 Message-ID: References: <20100307104252.GA19926@redhat.com> <40B76F19-1E35-4956-A34E-0A83A403F97D@suse.de> <20100307110457.GB20004@redhat.com> Mime-Version: 1.0 (Apple Message framework v1077) Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20100307110457.GB20004@redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: "Michael S. Tsirkin" Cc: Shirley Ma , Jamie Lokier , linux-kernel@vger.kernel.org, virtualization@lists.linux-foundation.org, Anthony Liguori , Amit Shah , Linus Torvalds , hch@lst.de List-Id: virtualization@lists.linuxfoundation.org On 07.03.2010, at 12:04, Michael S. Tsirkin wrote: > On Sun, Mar 07, 2010 at 11:57:23AM +0100, Alexander Graf wrote: >> >> On 07.03.2010, at 11:42, Michael S. Tsirkin wrote: >> >>> Linus, >>> Please pull another virtio fix: this one fixes hotplug and >>> was supposed to be applied by Rusty already >>> (http://lkml.org/lkml/2009/11/30/451), but apparently got lost, >>> somehow. >>> >>> Thanks! >>> >>> The following changes since commit 3119815912a220bdac943dfbdfee640414c0c611: >>> Michael S. Tsirkin (1): >>> virtio: fix out of range array access >>> >>> are available in the git repository at: >>> >>> git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git for-linus >>> >>> Michael S. Tsirkin (1): >>> virtio: set pci bus master enable bit >> >> This is still missing an identifier for the hypervisor. We need to know if the guest will ever set bus master enabled or not, so we can have a compat hack in qemu to not check for it. >> >> >> Alex > > This patch fixes hotplug with qemu 0.11-0.12 which can't make use of a new > identifier without a time machine. For 0.12 there's -stable and I don't think the bad patch is in 0.11 yet? > IMHO we can't add new identifiers in time for 2.6.34 anyway, so if you > want to do that, please send a patch and have it queued up for the next > release. But I don't think we need an identifier for compat hack. > Hypervisor can detect buggy guests when they set status to OK without > enabling bus master first. > > Makes sense? Ah, so you'd do the compat hack purely in qemu? Sounds good to me. Nevermind the noise - please take that patch :-). Alex