I saw similar issues on an older AMD (Opteron) box. With retpoline4 it was a hang but when I used a test version of what we currently have on master-next I got a reboot as well. The fault seems to happen very quickly, so there is not even output on a serial console.
Checking /sys/devices/system/cpu/vulnerabilities/spectre_v2 I got (not sure I remember the exact phrase) "full AMD retpoline". Then I tried booting with spectre_v2=off. That allows the VM guest to boot, so this can be isolated to the AMD spectre_v2 mitigation. -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1747507 Title: starting VMs causing retpoline4 to reboot Status in linux package in Ubuntu: Confirmed Bug description: Using 4.13.0-33.36~retpoline4 kernel test kernel from ppa https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/spectre/. CPU: AMD Ryzen 7 1700 Eight-Core Processor Ubuntu: 17.10. Steps to reproduce (has occurred both times I tried it): 1. Just run kvm with no options (or virsh start <> or start VM from GUI) Machine reboots Going back to 4.13.0-32 has virsh working again. qemu-system-x86_64 command works fine unless you use the enable kvm command. To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1747507/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
