On Mon, Oct 30, 2017 at 09:34:00AM +0100, Paolo Bonzini wrote: > On 27/10/2017 10:38, Fabian Grünbichler wrote: > > On Wed, Oct 25, 2017 at 11:07:54AM +0200, Fabian Grünbichler wrote: > >> On Wed, Oct 25, 2017 at 10:40:11AM +0200, Paolo Bonzini wrote: > >>> On 25/10/2017 10:12, Fabian Grünbichler wrote: > >>>> any update on this? 4.14 is soon out the door, and 4.13 with the > >>>> breaking change just got released as part of Ubuntu 17.10 / Artful.. > >>> > >>> I had posted a patch on October 10th, but no one replied with test results. > >>> > >> > >> sorry, missed the (second) thread. > >> > >> currently building, will get back with test results ASAP. > > > > so, while we haven't found any affected hardware in our test lab (not > > even in the to-be-discarded pile), feedback from users has been positive > > across the board for the following CPUs which were not able to use KVM > > without the patch: > > > > - Intel Xeon 3040 (Conroe) > > - Intel Xeon 5050 (Dempsey) > > - Intel Xeon 5130 (Woodcrest) > > - Intel Xeon E3-1256L V2 (Ivy Bridge) > > Ivy Bridge definitely worked without the patch. > > Can you send me a link to the exact patch you applied (pastebin or > something like that)? > the exact wording was the following, so it is possible that that user confused an unrelated (kernel or other) issue with the one at hand: -----%<----- After rolling back the 5.1 update, and staying on pve 5.0, testing the kernel posted above seams to work. However, I will be staying with the 4.10.17 kernel for now (and pve 5.0). -----%<----- the switch from 4.10 based to 4.13 based coincides with a minor release unfortunately. like I said, we don't have any test HW available, so we have to rely on user reports. there has been no further feedback so far. patch is available at http://sprunge.us/ULML
