On Tue, 28 Apr 2020 at 02:42, Sean Christopherson
<sean.j.christopherson@xxxxxxxxx> wrote:
>
> On Fri, Apr 24, 2020 at 02:22:44PM +0800, Wanpeng Li wrote:
> > From: Wanpeng Li <wanpengli@xxxxxxxxxxx>
> >
> > This patch implements handle preemption timer fastpath, after timer fire
> > due to VMX-preemption timer counts down to zero, handle it as soon as
> > possible and vmentry immediately without checking various kvm stuff when
> > possible.
> >
> > Testing on SKX Server.
> >
> > cyclictest in guest(w/o mwait exposed, adaptive advance lapic timer is default -1):
> >
> > 5540.5ns -> 4602ns 17%
> >
> > kvm-unit-test/vmexit.flat:
> >
> > w/o avanced timer:
> > tscdeadline_immed: 2885 -> 2431.25 15.7%
> > tscdeadline: 5668.75 -> 5188.5 8.4%
> >
> > w/ adaptive advance timer default -1:
> > tscdeadline_immed: 2965.25 -> 2520 15.0%
> > tscdeadline: 4663.75 -> 4537 2.7%
> >
> > Tested-by: Haiwei Li <lihaiwei@xxxxxxxxxxx>
> > Cc: Haiwei Li <lihaiwei@xxxxxxxxxxx>
> > Signed-off-by: Wanpeng Li <wanpengli@xxxxxxxxxxx>
> > ---
> > arch/x86/kvm/vmx/vmx.c | 16 ++++++++++++++++
> > 1 file changed, 16 insertions(+)
> >
> > diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> > index d21b66b..028967a 100644
> > --- a/arch/x86/kvm/vmx/vmx.c
> > +++ b/arch/x86/kvm/vmx/vmx.c
> > @@ -6560,12 +6560,28 @@ void vmx_update_host_rsp(struct vcpu_vmx *vmx, unsigned long host_rsp)
> > }
> > }
> >
> > +static enum exit_fastpath_completion handle_fastpath_preemption_timer(struct kvm_vcpu *vcpu)
>
> Somewhat offtopic, would it make sense to add a fastpath_t typedef? These
> enum lines are a bit long...
>
> > +{
> > + struct vcpu_vmx *vmx = to_vmx(vcpu);
> > +
> > + if (!vmx->req_immediate_exit &&
> > + !unlikely(vmx->loaded_vmcs->hv_timer_soft_disabled)) {
>
> Bad indentation.
>
> Also, this is is identical to handle_preemption_timer(), why not something
> like:
>
> static bool __handle_preemption_timer(struct vcpu)
> {
> struct vcpu_vmx *vmx = to_vmx(vcpu);
>
> if (!vmx->req_immediate_exit &&
> !unlikely(vmx->loaded_vmcs->hv_timer_soft_disabled)) {
> kvm_lapic_expired_hv_timer(vcpu);
> return true;
> }
>
> return false;
> }
>
> static enum exit_fastpath_completion handle_fastpath_preemption_timer(struct kvm_vcpu *vcpu)
> {
> if (__handle_preemption_timer(vcpu))
> return EXIT_FASTPATH_CONT_RUN;
> return EXIT_FASTPATH_NONE;
> }
>
> static int handle_preemption_timer(struct kvm_vcpu *vcpu)
> {
> __handle_preemption_timer(vcpu);
> return 1;
> }
>
Great! Thanks for making this nicer.
Wanpeng
