On Wed, Dec 12, 2018 at 10:53:35AM +0000, Suzuki K Poulose wrote:
>
>
> On 12/12/2018 10:29, Andrew Murray wrote:
> > Enable/disable event counters as appropriate when entering and exiting
> > the guest to enable support for guest or host only event counting.
> >
> > For both VHE and non-VHE we switch the counters between host/guest at
> > EL2. EL2 is filtered out by the PMU when we are using the :G modifier.
> >
> > The PMU may be on when we change which counters are enabled however
> > we avoid adding an isb as we instead rely on existing context
> > synchronisation events: the isb in kvm_arm_vhe_guest_exit for VHE and
> > the eret from the hvc in kvm_call_hyp.
> >
> > Signed-off-by: Andrew Murray <andrew.murray@xxxxxxx>
> > ---
> > arch/arm64/kvm/hyp/switch.c | 52 +++++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 52 insertions(+)
> >
> > diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
> > index d496ef5..9732ef7 100644
> > --- a/arch/arm64/kvm/hyp/switch.c
> > +++ b/arch/arm64/kvm/hyp/switch.c
> > @@ -373,6 +373,46 @@ static bool __hyp_text __hyp_switch_fpsimd(struct kvm_vcpu *vcpu)
> > return true;
> > }
> > +static bool __hyp_text __pmu_switch_to_guest(struct kvm_cpu_context *host_ctxt)
> > +{
> > + struct kvm_host_data *host;
> > + struct kvm_pmu_events *pmu;
> > + u32 clr, set;
> > +
> > + host = container_of(host_ctxt, struct kvm_host_data, host_ctxt);
> > + pmu = &host->pmu_events;
> > +
> > + clr = pmu->events_host & ~pmu->events_guest;
> > + set = pmu->events_guest & ~pmu->events_host;
>
> It may make sense to add in a comment explaining why we only set:
>
> events_guest & ~events_host
>
Yes I'll add a comment, especially as I've just spend 5 minutes trying to
remember why I did this.
Instead of assigning 'pmu->events_guest & ~pmu->events_host' to 'set' we
could have just assigned 'pmu->events_guest'. However consider the scenario
where an event is enabled for both host and guest - this would have resulted
in us always writing the (already set) bit to the system register. Therefore
with my approach we can potentially avoid any system register writes on a
hypervisor switch when the only events enabled are those for both host and
guest.
Thanks,
Andrew Murray
> Either way:
>
> Reviewed-by: Suzuki K Poulose <suzuki.poulose@xxxxxxx>
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
