timekeeping regression?

[Greg Troxel <gdt%lexort.com@localhost>]

I have an old dom0 box that I am running NetBSD 10 and Xen 4.18 on.
It's been mostly fine.

I update to RC3 (13 days ago says uptime).  I am seeing the dom0 not
really converged with ntp, offsets -5000 to -15000 (so 5 to 15s fast).

  kern.timecounter.choice = clockinterrupt(q=0, f=100 Hz) xen_system_time(q=10000, f=1000000000 Hz) ichlpcib0(q=1000, f=3579545 Hz) hpet0(q=2000, f=14318180 Hz) ACPI-Safe(q=900, f=3579545 Hz) dummy(q=-1000000, f=1000000 Hz)
  kern.timecounter.hardware = xen_system_time

There is an n-9 amd64 domU which seems pretty well synchronized.

  kern.timecounter.choice = xen_system_time(q=10000, f=1000000000 Hz) clockinterrupt(q=0, f=100 Hz) dummy(q=-1000000, f=1000000 Hz)
  kern.timecounter.hardware = xen_system_time

There is a n-9 i386 domU running pv shim:

  type="pvh"
  pvshim=1

and timekeeping is so bad I sometimes can't log in.  ntpq -p shows
265-365 seconds slow right now.

  kern.timecounter.choice = xen_system_time(q=10000, f=1000000000 Hz) clockinterrupt(q=0, f=100 Hz) dummy(q=-1000000, f=1000000 Hz)
  kern.timecounter.hardware = xen_system_time


Are others seeing this?