On Thu, Apr 21, 2016 at 08:01:58PM +0300, Roman Kagan wrote: > As to the drift problem, I reliably reproduce it, too, but it's a bug > somewhere in kvm_clock, and the new hyperv ref tsc page now demonstrates > it by virtue of relying on kvm_clock's values. On my machine I observe > as much as +14 ppm and I'll try to chase it down, but this is > independent of the hyperv ref tsc page patchset. A simple demonstration of it: in a CentOS 7.2 x86_64 guest: # dmesg | grep -Fw tsc: [ 0.000000] tsc: Detected 3411.482 MHz processor [ 2.192473] tsc: Refined TSC clocksource calibration: 3411.532 MHz The first value is derived from the kvm_clock's tsc_to_system_mul and tsc_shift, and matches hosts's vcpu->hw_tsc_khz. The second is calibrated using emulated HPET. The difference is those +14 ppm. This is on i7-2600, invariant TSC present, TSC scaling not present. I'll dig further but I'd appreciate any comment on whether it was within tolerance or not. Roman. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
