Re: [qemu patch 2/2] kvmclock: reduce kvmclock difference on migration

From: Paolo Bonzini <pbonzini@redhat.com>
To: Eduardo Habkost <ehabkost@redhat.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>,
	kvm@vger.kernel.org, qemu-devel@nongnu.org,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Juan Quintela <quintela@redhat.com>,
	Radim Krcmar <rkrcmar@redhat.com>
Subject: Re: [qemu patch 2/2] kvmclock: reduce kvmclock difference on migration
Date: Mon, 28 Nov 2016 16:12:55 +0100	[thread overview]
Message-ID: <44a6702d-e91e-2815-1e2e-aef67d750b9e@redhat.com> (raw)
In-Reply-To: <20161128142855.GC14328@thinpad.lan.raisama.net>

On 28/11/2016 15:28, Eduardo Habkost wrote:
> > +    s->src_use_reliable_get_clock = data.flags & KVM_CLOCK_TSC_STABLE;
> 
> I still don't understand the reasoning behind
> kvm_has_adjust_clock_stable() vs (flags & KVM_CLOCK_TSC_STABLE),
> but on either case, updating src_use_reliable_get_clock inside
> kvm_get_clock() looks like the right thing to do.

There are three possibility: the kernel tells you the clock is stable,
the kernel tells you the clock is unstable, the kernel is too old and
doesn't tell you anything.  Then:

    kvm_has_adjust_clock_stable() == true:
	if the clock is stable, KVM_CLOCK_TSC_STABLE will be set in "flags"
	if the clock is unstable, KVM_CLOCK_TSC_STABLE will be unset

    kvm_has_adjust_clock_stable() == false:
	you cannot know if the clock is stable

Paolo