First of all, I have to mention OpenBSD guests work perfectly on 6.7amd64 host even without previously discovered clock skew issue. I'm using tsc and ntpd to adjust guests' clock precisely. No any huge clock skew is detected. Good job!
It seems tsc does not help on linux guests. Clock is two times slower than host system's time. I didn't find a solution to solve it. Any suggestions? Right after update amd64 6.6->6.7 vmd machines unexpectedly hangs with linux guests. Tested with Debian 5.2.0 kernel and Alpine 5.4.43-1-virt kernel. Machines run one by one (only one vm is run in a time of hang). CPU utilization continuously increases from 20-25% when normal run up to 160-188% when guest hangs, according to host system's top. I tried to stop the vmd guest by 'vmctl stop linux' and CPU utilization decreased to ~99-100% but machine don't stop finally without additionally sending 'kill -HUP PID' signal to vmd process. I didn't see it in 6.6 with same linux guests. Some errors from guests boot process which can be relevant: Loading Linux 5.2.0-debian-amd64 Loading initial ramdist ... ACPI BIOS Error (bug): A valid RSDP was not found (20190509/tbxfroot-210) [Firmware Bug]: TSC doesn't count with P0 frequency! [Firmware Bug]: cpu 0, invalid IBS interrupt offset 0 (MSRC001103A=0x0000000000000000) mce: Unambe to init MCE device (rc:-5) tsc: Fast TSC calibration failed ... Loading Linux 5.4.43-1-virt #Alpine SMP [Firmware Bug]: TSC doesn't count with P0 frequency! tsc: Fast TSC calibration failed tsc: Unable to calibrate against PIT tsc: No referece (HPET/PMTIMER) available tsc: Marking TSC unstable due to could not calculate TSC khz Martin
