From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: msitranslate: question about qemu-upstream and pv support Date: Mon, 4 Aug 2014 10:26:05 -0400 Message-ID: <20140804142605.GG18425@laptop.dumpdata.com> References: <53D185B8.3000505@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Content-Disposition: inline In-Reply-To: <53D185B8.3000505@gmail.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Andrew Sorensen Cc: xen-devel@lists.xen.org List-Id: xen-devel@lists.xenproject.org On Thu, Jul 24, 2014 at 03:16:24PM -0700, Andrew Sorensen wrote: > I noticed that the patches for msitranslate do not seem to have made it > into qemu-upstream. The logic still exists in > xen/drivers/passthrough/io.c, though documentation suggests MSI-INTx > translation is only implemented for HVM guests. > > Based on the disabling of MSI-INTx by default, it seems safe to assume > that MSI translation is only supported under qemu-xen-traditional, and > only for HVM guests. Is this correct? My recollection is that it is an optimization. That is - if we do PCI passthrough - and the device is doing legacy interrpupts - we want to pass that (as it means updating QEMU) and instead doing an MSI - so that QEMU is not scheduled in. But most network cards don't use legacy interrupts. Perhaps the video cards use it - but for example the i915 is doing MSI and I would think that radeon and nouvau are moving that way too? > > I don't see any mention of this on this wiki page: > http://wiki.xen.org/wiki/QEMU_Upstream > #Missing_feature_from_the_good_old_qemu-dm That looks to be an omission. Do you have an account on the Wiki to update it? Are you Ok updating it? Thank you! > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xen.org > http://lists.xen.org/xen-devel