On 3/24/20 8:30 AM, Jim Fehlig wrote:
On 3/24/20 3:44 AM, Jiri Denemark wrote:
The situation on the first machine is a bit strange as there are no
features disabled in host-model CPU definition, which makes it unclear
why QEMU reports Cascadelake-Server as unusable (QEMU reports the
reason, but we don't do so yet).
It's not clear to me even after looking at the output of cpu-gather.
Anyway, would you mind running the tests/cputestdata/cpu-gather.sh
script on both machines (make sure to install qemu, python3, and cpuid
packages first) and send us the output so that we can check the CPU
models are properly detected?
Attached.
While experimenting with some guests on the first machine (kernel 4.12.14, qemu
3.1.1, libvirt.git master) I encountered
https://bugs.launchpad.net/intel/+bug/1828495 while trying to restore a guest
whose CPU was set to host-model. The guest starts fine, although I do see
qemu-system-x86_64: warning: host doesn't support requested feature:
CPUID.07H:ECX [bit 4]
in the guest log file. Save works as expected but then restore fails with
error: Failed to restore domain from /home/jim/sles12sp5-kvm.save
error: operation failed: guest CPU doesn't match specification: missing
features: ospke
I wouldn't be surprised if my kernel and qemu are missing some commits mentioned
in the LP bug (there are a lot of them!), but shouldn't it be possible to
restore the guest if libvirt previously started it successfully?
Regards,
Jim
