[PATCH v2 5/5] KVM: VMX: Handle preemption timer fastpath

[Wanpeng Li <kernellwp@gmail.com>]

From: Wanpeng Li <wanpengli@tencent.com>

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):

5632.75ns -> 4559.25ns, 19%

kvm-unit-test/vmexit.flat:

w/o APICv, w/o advance timer:
tscdeadline_immed: 4780.75 -> 3851    19.4%
tscdeadline:       7474    -> 6528.5  12.7%

w/o APICv, w/ adaptive advance timer default -1:
tscdeadline_immed: 4845.75 -> 3930.5  18.9%
tscdeadline:       6048    -> 5871.75    3%

w/ APICv, w/o avanced timer:
tscdeadline_immed: 2919    -> 2467.75 15.5%
tscdeadline:       5661.75 -> 5188.25  8.4%

w/ APICv, w/ adaptive advance timer default -1:
tscdeadline_immed: 3018.5  -> 2561    15.2%
tscdeadline:       4663.75 -> 4537     2.7%

Tested-by: Haiwei Li <lihaiwei@tencent.com>
Cc: Haiwei Li <lihaiwei@tencent.com>
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
---
 arch/x86/kvm/lapic.c   | 19 +++++++++++++++++++
 arch/x86/kvm/lapic.h   |  1 +
 arch/x86/kvm/vmx/vmx.c | 22 ++++++++++++++++++++++
 3 files changed, 42 insertions(+)

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index d652bd9..2741931 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -1899,6 +1899,25 @@ void kvm_lapic_expired_hv_timer(struct kvm_vcpu *vcpu)
 EXPORT_SYMBOL_GPL(kvm_lapic_expired_hv_timer);
 
 static void kvm_inject_apic_timer_irqs_fast(struct kvm_vcpu *vcpu);
+bool kvm_lapic_expired_hv_timer_fast(struct kvm_vcpu *vcpu)
+{
+	struct kvm_lapic *apic = vcpu->arch.apic;
+	struct kvm_timer *ktimer = &apic->lapic_timer;
+
+	if (!apic_lvtt_tscdeadline(apic) ||
+		!ktimer->hv_timer_in_use ||
+		atomic_read(&ktimer->pending))
+		return 0;
+
+	WARN_ON(swait_active(&vcpu->wq));
+	cancel_hv_timer(apic);
+
+	ktimer->expired_tscdeadline = ktimer->tscdeadline;
+	kvm_inject_apic_timer_irqs_fast(vcpu);
+
+	return 1;
+}
+EXPORT_SYMBOL_GPL(kvm_lapic_expired_hv_timer_fast);
 
 void kvm_lapic_switch_to_hv_timer(struct kvm_vcpu *vcpu)
 {
diff --git a/arch/x86/kvm/lapic.h b/arch/x86/kvm/lapic.h
index 5ef1364..1b5abd8 100644
--- a/arch/x86/kvm/lapic.h
+++ b/arch/x86/kvm/lapic.h
@@ -252,6 +252,7 @@ bool kvm_lapic_hv_timer_in_use(struct kvm_vcpu *vcpu);
 void kvm_lapic_restart_hv_timer(struct kvm_vcpu *vcpu);
 bool kvm_can_post_timer_interrupt(struct kvm_vcpu *vcpu);
 int kvm_set_lapic_tscdeadline_msr_fast(struct kvm_vcpu *vcpu, u64 data);
+bool kvm_lapic_expired_hv_timer_fast(struct kvm_vcpu *vcpu);
 
 static inline enum lapic_mode kvm_apic_mode(u64 apic_base)
 {
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index 2613e58..527d1c1 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -6569,12 +6569,34 @@ void vmx_update_host_rsp(struct vcpu_vmx *vmx, unsigned long host_rsp)
 	}
 }
 
+static void vmx_cancel_hv_timer(struct kvm_vcpu *vcpu);
+
+static enum exit_fastpath_completion handle_fastpath_preemption_timer(struct kvm_vcpu *vcpu)
+{
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
+
+	if (kvm_need_cancel_enter_guest(vcpu) ||
+		kvm_event_needs_reinjection(vcpu))
+		return EXIT_FASTPATH_NONE;
+
+	if (!vmx->req_immediate_exit &&
+		!unlikely(vmx->loaded_vmcs->hv_timer_soft_disabled) &&
+		kvm_lapic_expired_hv_timer_fast(vcpu)) {
+		trace_kvm_exit(EXIT_REASON_PREEMPTION_TIMER, vcpu, KVM_ISA_VMX);
+		return EXIT_FASTPATH_CONT_RUN;
+	}
+
+	return EXIT_FASTPATH_NONE;
+}
+
 static enum exit_fastpath_completion vmx_exit_handlers_fastpath(struct kvm_vcpu *vcpu)
 {
 	if (!is_guest_mode(vcpu)) {
 		switch (to_vmx(vcpu)->exit_reason) {
 		case EXIT_REASON_MSR_WRITE:
 			return handle_fastpath_set_msr_irqoff(vcpu);
+		case EXIT_REASON_PREEMPTION_TIMER:
+			return handle_fastpath_preemption_timer(vcpu);
 		default:
 			return EXIT_FASTPATH_NONE;
 		}
-- 
2.7.4