From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:50524) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b3m2x-0000W3-SM for qemu-devel@nongnu.org; Fri, 20 May 2016 11:09:04 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1b3m2u-0003IC-KB for qemu-devel@nongnu.org; Fri, 20 May 2016 11:09:03 -0400 Received: from mx1.redhat.com ([209.132.183.28]:40160) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1b3m2u-0003I5-ER for qemu-devel@nongnu.org; Fri, 20 May 2016 11:09:00 -0400 Date: Fri, 20 May 2016 09:08:57 -0600 From: Alex Williamson Message-ID: <20160520090857.22174bd9@t450s.home> In-Reply-To: <1463746759.3526.88.camel@redhat.com> References: <20160517201613.6175.45578.stgit@gimli.home> <20160517144213.2c54d33e@t450s.home> <1463581489.30045.68.camel@redhat.com> <20160518124523.506dc1a3@t450s.home> <1463746759.3526.88.camel@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] vfio IGD assignment (was Re: [PATCH v6 0/8] Series short description) List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Gerd Hoffmann Cc: allen.m.kay@intel.com, qemu-devel@nongnu.org, kvm@vger.kernel.org On Fri, 20 May 2016 14:19:19 +0200 Gerd Hoffmann wrote: > On Mi, 2016-05-18 at 12:45 -0600, Alex Williamson wrote: > > On Wed, 18 May 2016 16:24:49 +0200 > > Gerd Hoffmann wrote: > > > > > Hi, > > > > > > > I believe we're no longer blocked by SeaBIOS being willing to > > > > consume the new fw_cfg entries, so barring feedback otherwise, my > > > > plan will be to send a pull request including these changes (after > > > > an appropriate review period), ask Kevin to apply the matching > > > > SeaBIOS change, then see if Gerd will help me pull that support > > > > into the default QEMU SeaBIOS image. > > > > > > Sounds good. Patches look sane to me too. > > > > > > Will try to set aside some time to test them later this week (and also > > > push a branch with updated seabios etc for others to try). > > > > Thanks Gerd! > > Works flawlessly. > > Linux works fine with the igd assigned as primary (pc machine type and > igd using slot 2). > > Linux doesn't work in UPT mode for me. That is more a guest driver > issue though. The i915 driver of older kernels Oopses. The i915 driver > of recent kernels (4.4.5+) loads fine on q35, but fails to drive my > monitor, no matter whenever I enable the opregion or not. For pc an > additional kernel patch is needed (which hopefully lands upstream soon) > to reach feature parity (as-in: no Oops on load) with q35. Yes, I've never had success with UPT for Linux guests either. > For windows UPT mode is the only thing supported by the drivers. Needs > the opregion enabled to show something on the connected display. Works > fine on both pc and q35, in parallel with a emulated display, and you > can configure things as you want (mirror, expand desktop, use only one > of the displays). UPT is probably the only thing Intel cares to support on Windows, yes, but did you have any issues or try legacy mode on Windows? I find that Windows 7/8/10 all work well with legacy mode, including the basic drivers in the shrink-wrap images. Please let me know if you find otherwise. > > Reviewed-by: Gerd Hoffmann > Tested-by: Gerd Hoffmann Awesome, thanks! Alex