On Tue, 2018-09-25 at 11:18 +0300, Sakari Ailus wrote: > Hi Junaid, > > On Tue, Sep 25, 2018 at 01:11:06AM -0700, Junaid Shahid wrote: > > > > On 09/24/2018 12:57 PM, Sakari Ailus wrote: > > > > > > > > > I'm not subscribed to the kvm list nor I did find the original e-mail with > > > the headers so I constructed the reply from the bits I did find. If > > > something goes wrong it's probably because of that. > > > > > > This patch in the host kernel breaks KVM guests on Core 2 Duo (T9500). KVM > > > guests do not boot on plain v4.18.9 (which contains the patch) host; > > > reverting this patch fixes it. > > Hi Sakari, > > > > Could you please let me know what OS is running in the guest? Also, is > > there some kernel panic etc. in the guest OS when it doesn't boot, or > > does it fail in some other way? > The guest is Debian running v4.18.9 kernel. > > Everything simply stops once I'd expect the kernel to start executing the > initrd scripts. Before figuring out the host kernel was at fault, I was > first wondering if there was something wrong with the guest itself and > disabled a few things such as APIC or HPET, and different oopses followed > from that. > > If there's no immediate, obvious fix, I'd strongly suggest to revert the > patch. I was able to reproduce this by disabling EPT on modern hardware. I have a fix, I'll send a patch shortly.
