From: Anthony Liguori Subject: Re: Adding an IPMI BMC device to KVM Date: Sun, 06 May 2012 09:35:10 -0500 Message-ID: <4FA68C1E.3070503@codemonkey.ws> References: <4FA429BA.3040006@acm.org> <4FA6788A.8080500@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: qemu-devel , Corey Minyard , minyard@acm.org, kvm@vger.kernel.org To: Avi Kivity Return-path: In-Reply-To: <4FA6788A.8080500@redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+gceq-qemu-devel=gmane.org@nongnu.org Sender: qemu-devel-bounces+gceq-qemu-devel=gmane.org@nongnu.org List-Id: kvm.vger.kernel.org On 05/06/2012 08:11 AM, Avi Kivity wrote: > (copied qemu-devel) > > On 05/04/2012 10:10 PM, Corey Minyard wrote: >> >> Either way, is this interesting for including into KVM? > > Not kvm, but certainly it would make a good addition to qemu, which kvm > then uses. > >> Does anyone have any >> opinions on the possible ways to implement this? > > My preference would be the second alternative. The issue you raise is a > good one. There are two ways we can approach it: > > - have the management system intercept IPMI requests, start up a qemu > instance (if it's down), and let it handle the event. > - change the whole system to keep a running qemu even when the guest is > down. This is a much larger change; it involves reducing the memory > footprint to almost nothing when the guest is down (deallocating memory > and threads) so it doesn't impact guest density, but it allows for other > minor features such as wake-on-LAN and RTC alarm wakeups. libvirt is essentially the BMC for a virtual guest. I would suggest looking at implementing an IPMI interface to libvirt and exposing it to the guest through a USB RNDIS device. Regards, Anthony Liguori >