From: David Woodhouse <dwmw@xxxxxxxxxxxx>
Most of the time there's no need to kick the vCPU and deliver the timer
event through kvm_xen_inject_timer_irqs(). Use kvm_xen_set_evtchn_fast()
directly from the timer callback, and only fall back to the slow path
when it's necessary to do so.
This gives a significant improvement in timer latency testing (using
nanosleep() for various periods and then measuring the actual time
elapsed).
However, there was a reason¹ the fast path was dropped when this support
was first added. The current code holds vcpu->mutex for all operations on
the kvm->arch.timer_expires field, and the fast path introduces potential
race conditions. So... ensure the hrtimer is *cancelled* before making
changes in kvm_xen_start_timer(), and also when reading the values out
for KVM_XEN_VCPU_ATTR_TYPE_TIMER.
Add some sanity checks to ensure the truth of the claim that all the
other code paths are run with the vcpu loaded. And use hrtimer_cancel()
directly from kvm_xen_destroy_vcpu() to avoid a false positive from the
check in kvm_xen_stop_timer().
¹ https://lore.kernel.org/kvm/846caa99-2e42-4443-1070-84e49d2f11d2@xxxxxxxxxx/
Signed-off-by: David Woodhouse <dwmw@xxxxxxxxxxxx>
---
• v2: Remember, and deal with, those races.
arch/x86/kvm/xen.c | 64 +++++++++++++++++++++++++++++++++++++++++-----
1 file changed, 58 insertions(+), 6 deletions(-)