Page MenuHomeFreeBSD

paravirt: Add KVM clock timecounter support
AbandonedPublic

Authored by bryanv on Jan 4 2015, 5:57 AM.

Details

Reviewers
mmacy
jhb
Summary

This allows a KVM guest to get an accurate cycle counter without having to do a VM exit, which happens when either HPET or ACPI is used.

KVM clock also supports getting the wall clock time but that is not implemented yet.

Diff Detail

Lint
Lint Skipped
Unit
Unit Tests Skipped

Event Timeline

bryanv updated this revision to Diff 2988.Jan 4 2015, 5:57 AM
bryanv retitled this revision from to paravirt: Add KVM clock timecounter support.
bryanv updated this object.
bryanv edited the test plan for this revision. (Show Details)
bryanv set the repository for this revision to rS FreeBSD src repository.
bryanv updated this object.

What's the status of this patch. It looks very promising.
In fact I was writing kvmclock support code myself until I found this one. I think it is much better than the old tsc code.

sys/x86/include/kvm.h
44

in fact, kvm paravirt detection stays identical with x2apic support detection in linux source tree.

What's the status of this patch. It looks very promising.
In fact I was writing kvmclock support code myself until I found this one. I think it is much better than the old tsc code.

I lost my spare FreeBSD before I was able to get this and prerequisites changes into the tree. I'm trying to resume up my FreeBSD work and this is one of the changes I hope to get committed.

badger added a subscriber: badger.Dec 21 2016, 4:39 PM
swills added a subscriber: swills.Jul 8 2018, 9:02 PM
mmacy set the repository for this revision to rS FreeBSD src repository.
mmacy added a comment.Jul 8 2018, 10:17 PM

@bryanv What do you feel remains to be done?

bryanv added a subscriber: stevek.Jul 9 2018, 3:25 PM
In D1435#343209, @mmacy wrote:

@bryanv What do you feel remains to be done?

The change in this review depends on some hypervisor detection changes that I don't think make sense anymore. @stevek has done the work to refactor this change on to HEAD that I hope he's able to submit soon.

The change in this review depends on some hypervisor detection changes that I don't think make sense anymore. @stevek has done the work to refactor this change on to HEAD that I hope he's able to submit soon.

See D16305 and D16306

bryanv abandoned this revision.Jul 18 2018, 3:04 PM