On 2021-11-11 02:15, Thomas Munro wrote:
> I dunno. Clocks on virtualised systems and even metal seem to be a
> minefield of quirks and heuristics. Some discussion, may or may not
> be relevant:
>
> https://marc.info/?l=openbsd-tech&m=161657532610882&w=2
Just for fun I compiled and ran the test program and this was the result:
$ ./monotime
589047 Starting
284971 Starting
542819 Starting
315557 Starting
589047 Stopped
542819 Back 2728606.093473837 => 2728605.963205128
542819 Stopped
284971 Stopped
315557 Stopped
above was on 6.9
then I tried the same on my older 5.9 animal and this was the result:
./monotime
1003853 Starting
1004437 Starting
1032556 Starting
1001887 Starting
1004437 Stopped
1003853 Stopped
1032556 Stopped
1001887 Stopped
it's also running under the same VMWARE 6.7 instance / machine.
dmesg on that machine also indicates that it's using:
acpitimer0 at acpi0: 3579545 Hz, 24 bits
so looks like a kernel bug / regression in recent OpenBSD kernels then?
/Mikael
