KVM: x86/vPMU: Add lazy mechanism to release perf_event per vPMC

Currently, a host perf_event is created for a vPMC functionality emulation.
It’s unpredictable to determine if a disabled perf_event will be reused.
If they are disabled and are not reused for a considerable period of time,
those obsolete perf_events would increase host context switch overhead that
could have been avoided.

If the guest doesn't WRMSR any of the vPMC's MSRs during an entire vcpu
sched time slice, and its independent enable bit of the vPMC isn't set,
we can predict that the guest has finished the use of this vPMC, and then
do request KVM_REQ_PMU in kvm_arch_sched_in and release those perf_events
in the first call of kvm_pmu_handle_event() after the vcpu is scheduled in.

This lazy mechanism delays the event release time to the beginning of the
next scheduled time slice if vPMC's MSRs aren't changed during this time
slice. If guest comes back to use this vPMC in next time slice, a new perf
event would be re-created via perf_event_create_kernel_counter() as usual.

Suggested-by: Wei Wang <wei.w.wang@intel.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Like Xu <like.xu@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

6 files changed, 84 insertions, 0 deletions

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index a87a6c98adee..20bb2fc0883a 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -473,6 +473,20 @@ struct kvm_pmu {
 	struct kvm_pmc fixed_counters[INTEL_PMC_MAX_FIXED];
 	struct irq_work irq_work;
 	DECLARE_BITMAP(reprogram_pmi, X86_PMC_IDX_MAX);
+	DECLARE_BITMAP(all_valid_pmc_idx, X86_PMC_IDX_MAX);
+	DECLARE_BITMAP(pmc_in_use, X86_PMC_IDX_MAX);
+
+	/*
+	 * The gate to release perf_events not marked in
+	 * pmc_in_use only once in a vcpu time slice.
+	 */
+	bool need_cleanup;
+
+	/*
+	 * The total number of programmed perf_events and it helps to avoid
+	 * redundant check before cleanup if guest don't use vPMU at all.
+	 */
+	u8 event_count;
 };
 
 struct kvm_pmu_ops;
diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c
index 99565de5410a..d5e6d5b3f06f 100644
--- a/arch/x86/kvm/pmu.c
+++ b/arch/x86/kvm/pmu.c
@@ -135,6 +135,7 @@ static void pmc_reprogram_counter(struct kvm_pmc *pmc, u32 type,
 	}
 
 	pmc->perf_event = event;
+	pmc_to_pmu(pmc)->event_count++;
 	clear_bit(pmc->idx, pmc_to_pmu(pmc)->reprogram_pmi);
 }
 
@@ -304,6 +305,14 @@ void kvm_pmu_handle_event(struct kvm_vcpu *vcpu)
 
 		reprogram_counter(pmu, bit);
 	}
+
+	/*
+	 * Unused perf_events are only released if the corresponding MSRs
+	 * weren't accessed during the last vCPU time slice. kvm_arch_sched_in
+	 * triggers KVM_REQ_PMU if cleanup is needed.
+	 */
+	if (unlikely(pmu->need_cleanup))
+		kvm_pmu_cleanup(vcpu);
 }
 
 /* check if idx is a valid index to access PMU */
@@ -379,6 +388,15 @@ bool kvm_pmu_is_valid_msr(struct kvm_vcpu *vcpu, u32 msr)
 		kvm_x86_ops->pmu_ops->is_valid_msr(vcpu, msr);
 }
 
+static void kvm_pmu_mark_pmc_in_use(struct kvm_vcpu *vcpu, u32 msr)
+{
+	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
+	struct kvm_pmc *pmc = kvm_x86_ops->pmu_ops->msr_idx_to_pmc(vcpu, msr);
+
+	if (pmc)
+		__set_bit(pmc->idx, pmu->pmc_in_use);
+}
+
 int kvm_pmu_get_msr(struct kvm_vcpu *vcpu, u32 msr, u64 *data)
 {
 	return kvm_x86_ops->pmu_ops->get_msr(vcpu, msr, data);
@@ -386,6 +404,7 @@ int kvm_pmu_get_msr(struct kvm_vcpu *vcpu, u32 msr, u64 *data)
 
 int kvm_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 {
+	kvm_pmu_mark_pmc_in_use(vcpu, msr_info->index);
 	return kvm_x86_ops->pmu_ops->set_msr(vcpu, msr_info);
 }
 
@@ -413,9 +432,45 @@ void kvm_pmu_init(struct kvm_vcpu *vcpu)
 	memset(pmu, 0, sizeof(*pmu));
 	kvm_x86_ops->pmu_ops->init(vcpu);
 	init_irq_work(&pmu->irq_work, kvm_pmi_trigger_fn);
+	pmu->event_count = 0;
+	pmu->need_cleanup = false;
 	kvm_pmu_refresh(vcpu);
 }
 
+static inline bool pmc_speculative_in_use(struct kvm_pmc *pmc)
+{
+	struct kvm_pmu *pmu = pmc_to_pmu(pmc);
+
+	if (pmc_is_fixed(pmc))
+		return fixed_ctrl_field(pmu->fixed_ctr_ctrl,
+			pmc->idx - INTEL_PMC_IDX_FIXED) & 0x3;
+
+	return pmc->eventsel & ARCH_PERFMON_EVENTSEL_ENABLE;
+}
+
+/* Release perf_events for vPMCs that have been unused for a full time slice.  */
+void kvm_pmu_cleanup(struct kvm_vcpu *vcpu)
+{
+	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
+	struct kvm_pmc *pmc = NULL;
+	DECLARE_BITMAP(bitmask, X86_PMC_IDX_MAX);
+	int i;
+
+	pmu->need_cleanup = false;
+
+	bitmap_andnot(bitmask, pmu->all_valid_pmc_idx,
+		      pmu->pmc_in_use, X86_PMC_IDX_MAX);
+
+	for_each_set_bit(i, bitmask, X86_PMC_IDX_MAX) {
+		pmc = kvm_x86_ops->pmu_ops->pmc_idx_to_pmc(pmu, i);
+
+		if (pmc && pmc->perf_event && !pmc_speculative_in_use(pmc))
+			pmc_stop_counter(pmc);
+	}
+
+	bitmap_zero(pmu->pmc_in_use, X86_PMC_IDX_MAX);
+}
+
 void kvm_pmu_destroy(struct kvm_vcpu *vcpu)
 {
 	kvm_pmu_reset(vcpu);
diff --git a/arch/x86/kvm/pmu.h b/arch/x86/kvm/pmu.h
index 7eba298587dc..b7a625874203 100644
--- a/arch/x86/kvm/pmu.h
+++ b/arch/x86/kvm/pmu.h
@@ -62,6 +62,7 @@ static inline void pmc_release_perf_event(struct kvm_pmc *pmc)
 		perf_event_release_kernel(pmc->perf_event);
 		pmc->perf_event = NULL;
 		pmc->current_config = 0;
+		pmc_to_pmu(pmc)->event_count--;
 	}
 }
 
@@ -126,6 +127,7 @@ int kvm_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info);
 void kvm_pmu_refresh(struct kvm_vcpu *vcpu);
 void kvm_pmu_reset(struct kvm_vcpu *vcpu);
 void kvm_pmu_init(struct kvm_vcpu *vcpu);
+void kvm_pmu_cleanup(struct kvm_vcpu *vcpu);
 void kvm_pmu_destroy(struct kvm_vcpu *vcpu);
 int kvm_vm_ioctl_set_pmu_event_filter(struct kvm *kvm, void __user *argp);
 
diff --git a/arch/x86/kvm/pmu_amd.c b/arch/x86/kvm/pmu_amd.c
index e87d34136047..ce0b10fe5e2b 100644
--- a/arch/x86/kvm/pmu_amd.c
+++ b/arch/x86/kvm/pmu_amd.c
@@ -279,6 +279,7 @@ static void amd_pmu_refresh(struct kvm_vcpu *vcpu)
 	pmu->counter_bitmask[KVM_PMC_FIXED] = 0;
 	pmu->nr_arch_fixed_counters = 0;
 	pmu->global_status = 0;
+	bitmap_set(pmu->all_valid_pmc_idx, 0, pmu->nr_arch_gp_counters);
 }
 
 static void amd_pmu_init(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/vmx/pmu_intel.c b/arch/x86/kvm/vmx/pmu_intel.c
index 9b1ddc42f604..b5a16379f534 100644
--- a/arch/x86/kvm/vmx/pmu_intel.c
+++ b/arch/x86/kvm/vmx/pmu_intel.c
@@ -46,6 +46,7 @@ static void reprogram_fixed_counters(struct kvm_pmu *pmu, u64 data)
 		if (old_ctrl == new_ctrl)
 			continue;
 
+		__set_bit(INTEL_PMC_IDX_FIXED + i, pmu->pmc_in_use);
 		reprogram_fixed_counter(pmc, new_ctrl, i);
 	}
 
@@ -329,6 +330,11 @@ static void intel_pmu_refresh(struct kvm_vcpu *vcpu)
 	    (boot_cpu_has(X86_FEATURE_HLE) || boot_cpu_has(X86_FEATURE_RTM)) &&
 	    (entry->ebx & (X86_FEATURE_HLE|X86_FEATURE_RTM)))
 		pmu->reserved_bits ^= HSW_IN_TX|HSW_IN_TX_CHECKPOINTED;
+
+	bitmap_set(pmu->all_valid_pmc_idx,
+		0, pmu->nr_arch_gp_counters);
+	bitmap_set(pmu->all_valid_pmc_idx,
+		INTEL_PMC_MAX_GENERIC, pmu->nr_arch_fixed_counters);
 }
 
 static void intel_pmu_init(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 72189160bb81..8db7275d313f 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -9455,7 +9455,13 @@ void kvm_arch_vcpu_uninit(struct kvm_vcpu *vcpu)
 
 void kvm_arch_sched_in(struct kvm_vcpu *vcpu, int cpu)
 {
+	struct kvm_pmu *pmu = vcpu_to_pmu(vcpu);
+
 	vcpu->arch.l1tf_flush_l1d = true;
+	if (pmu->version && unlikely(pmu->event_count)) {
+		pmu->need_cleanup = true;
+		kvm_make_request(KVM_REQ_PMU, vcpu);
+	}
 	kvm_x86_ops->sched_in(vcpu, cpu);
 }