KVM: x86: interrupt based APF 'page ready' event delivery

Concerns were expressed around APF delivery via synthetic #PF exception as
in some cases such delivery may collide with real page fault. For 'page
ready' notifications we can easily switch to using an interrupt instead.
Introduce new MSR_KVM_ASYNC_PF_INT mechanism and deprecate the legacy one.

One notable difference between the two mechanisms is that interrupt may not
get handled immediately so whenever we would like to deliver next event
(regardless of its type) we must be sure the guest had read and cleared
previous event in the slot.

While on it, get rid on 'type 1/type 2' names for APF events in the
documentation as they are causing confusion. Use 'page not present'
and 'page ready' everywhere instead.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20200525144125.143875-6-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

3 files changed, 81 insertions, 28 deletions

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index f3897e417b69..2d39571451a0 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -767,7 +767,9 @@ struct kvm_vcpu_arch {
 		bool halted;
 		gfn_t gfns[ASYNC_PF_PER_VCPU];
 		struct gfn_to_hva_cache data;
-		u64 msr_val;
+		u64 msr_en_val; /* MSR_KVM_ASYNC_PF_EN */
+		u64 msr_int_val; /* MSR_KVM_ASYNC_PF_INT */
+		u16 vec;
 		u32 id;
 		bool send_user_only;
 		u32 host_apf_flags;
diff --git a/arch/x86/include/uapi/asm/kvm_para.h b/arch/x86/include/uapi/asm/kvm_para.h
index d1cd5c0f431a..1d37d616b1fc 100644
--- a/arch/x86/include/uapi/asm/kvm_para.h
+++ b/arch/x86/include/uapi/asm/kvm_para.h
@@ -50,6 +50,7 @@
 #define MSR_KVM_STEAL_TIME  0x4b564d03
 #define MSR_KVM_PV_EOI_EN      0x4b564d04
 #define MSR_KVM_POLL_CONTROL	0x4b564d05
+#define MSR_KVM_ASYNC_PF_INT	0x4b564d06
 
 struct kvm_steal_time {
 	__u64 steal;
@@ -81,6 +82,11 @@ struct kvm_clock_pairing {
 #define KVM_ASYNC_PF_ENABLED			(1 << 0)
 #define KVM_ASYNC_PF_SEND_ALWAYS		(1 << 1)
 #define KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT	(1 << 2)
+#define KVM_ASYNC_PF_DELIVERY_AS_INT		(1 << 3)
+
+/* MSR_KVM_ASYNC_PF_INT */
+#define KVM_ASYNC_PF_VEC_MASK			GENMASK(7, 0)
+
 
 /* Operations for KVM_HC_MMU_OP */
 #define KVM_MMU_OP_WRITE_PTE            1
@@ -112,8 +118,12 @@ struct kvm_mmu_op_release_pt {
 #define KVM_PV_REASON_PAGE_READY 2
 
 struct kvm_vcpu_pv_apf_data {
+	/* Used for 'page not present' events delivered via #PF */
 	__u32 flags;
-	__u32 token; /* Used for page ready notification only */
+
+	/* Used for 'page ready' events delivered via interrupt notification */
+	__u32 token;
+
 	__u8 pad[56];
 	__u32 enabled;
 };
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 0e79b37b2b7e..e6f3ec5193b2 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1248,7 +1248,7 @@ static const u32 emulated_msrs_all[] = {
 	HV_X64_MSR_TSC_EMULATION_STATUS,
 
 	MSR_KVM_ASYNC_PF_EN, MSR_KVM_STEAL_TIME,
-	MSR_KVM_PV_EOI_EN,
+	MSR_KVM_PV_EOI_EN, MSR_KVM_ASYNC_PF_INT,
 
 	MSR_IA32_TSC_ADJUST,
 	MSR_IA32_TSCDEADLINE,
@@ -2673,17 +2673,24 @@ out:
 	return r;
 }
 
+static inline bool kvm_pv_async_pf_enabled(struct kvm_vcpu *vcpu)
+{
+	u64 mask = KVM_ASYNC_PF_ENABLED | KVM_ASYNC_PF_DELIVERY_AS_INT;
+
+	return (vcpu->arch.apf.msr_en_val & mask) == mask;
+}
+
 static int kvm_pv_enable_async_pf(struct kvm_vcpu *vcpu, u64 data)
 {
 	gpa_t gpa = data & ~0x3f;
 
-	/* Bits 3:5 are reserved, Should be zero */
-	if (data & 0x38)
+	/* Bits 4:5 are reserved, Should be zero */
+	if (data & 0x30)
 		return 1;
 
-	vcpu->arch.apf.msr_val = data;
+	vcpu->arch.apf.msr_en_val = data;
 
-	if (!(data & KVM_ASYNC_PF_ENABLED)) {
+	if (!kvm_pv_async_pf_enabled(vcpu)) {
 		kvm_clear_async_pf_completion_queue(vcpu);
 		kvm_async_pf_hash_reset(vcpu);
 		return 0;
@@ -2695,7 +2702,25 @@ static int kvm_pv_enable_async_pf(struct kvm_vcpu *vcpu, u64 data)
 
 	vcpu->arch.apf.send_user_only = !(data & KVM_ASYNC_PF_SEND_ALWAYS);
 	vcpu->arch.apf.delivery_as_pf_vmexit = data & KVM_ASYNC_PF_DELIVERY_AS_PF_VMEXIT;
+
 	kvm_async_pf_wakeup_all(vcpu);
+
+	return 0;
+}
+
+static int kvm_pv_enable_async_pf_int(struct kvm_vcpu *vcpu, u64 data)
+{
+	/* Bits 8-63 are reserved */
+	if (data >> 8)
+		return 1;
+
+	if (!lapic_in_kernel(vcpu))
+		return 1;
+
+	vcpu->arch.apf.msr_int_val = data;
+
+	vcpu->arch.apf.vec = data & KVM_ASYNC_PF_VEC_MASK;
+
 	return 0;
 }
 
@@ -2917,6 +2942,10 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		if (kvm_pv_enable_async_pf(vcpu, data))
 			return 1;
 		break;
+	case MSR_KVM_ASYNC_PF_INT:
+		if (kvm_pv_enable_async_pf_int(vcpu, data))
+			return 1;
+		break;
 	case MSR_KVM_STEAL_TIME:
 
 		if (unlikely(!sched_info_on()))
@@ -3191,7 +3220,10 @@ int kvm_get_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		msr_info->data = vcpu->arch.time;
 		break;
 	case MSR_KVM_ASYNC_PF_EN:
-		msr_info->data = vcpu->arch.apf.msr_val;
+		msr_info->data = vcpu->arch.apf.msr_en_val;
+		break;
+	case MSR_KVM_ASYNC_PF_INT:
+		msr_info->data = vcpu->arch.apf.msr_int_val;
 		break;
 	case MSR_KVM_STEAL_TIME:
 		msr_info->data = vcpu->arch.st.msr_val;
@@ -9553,7 +9585,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event)
 	vcpu->arch.cr2 = 0;
 
 	kvm_make_request(KVM_REQ_EVENT, vcpu);
-	vcpu->arch.apf.msr_val = 0;
+	vcpu->arch.apf.msr_en_val = 0;
+	vcpu->arch.apf.msr_int_val = 0;
 	vcpu->arch.st.msr_val = 0;
 
 	kvmclock_reset(vcpu);
@@ -10430,10 +10463,22 @@ static inline int apf_put_user_notpresent(struct kvm_vcpu *vcpu)
 
 static inline int apf_put_user_ready(struct kvm_vcpu *vcpu, u32 token)
 {
-	u64 val = (u64)token << 32 | KVM_PV_REASON_PAGE_READY;
+	unsigned int offset = offsetof(struct kvm_vcpu_pv_apf_data, token);
 
-	return kvm_write_guest_cached(vcpu->kvm, &vcpu->arch.apf.data, &val,
-				      sizeof(val));
+	return kvm_write_guest_offset_cached(vcpu->kvm, &vcpu->arch.apf.data,
+					     &token, offset, sizeof(token));
+}
+
+static inline bool apf_pageready_slot_free(struct kvm_vcpu *vcpu)
+{
+	unsigned int offset = offsetof(struct kvm_vcpu_pv_apf_data, token);
+	u32 val;
+
+	if (kvm_read_guest_offset_cached(vcpu->kvm, &vcpu->arch.apf.data,
+					 &val, offset, sizeof(val)))
+		return false;
+
+	return !val;
 }
 
 static bool kvm_can_deliver_async_pf(struct kvm_vcpu *vcpu)
@@ -10441,9 +10486,8 @@ static bool kvm_can_deliver_async_pf(struct kvm_vcpu *vcpu)
 	if (!vcpu->arch.apf.delivery_as_pf_vmexit && is_guest_mode(vcpu))
 		return false;
 
-	if (!(vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED) ||
-	    (vcpu->arch.apf.send_user_only &&
-	     kvm_x86_ops.get_cpl(vcpu) == 0))
+	if (!kvm_pv_async_pf_enabled(vcpu) ||
+	    (vcpu->arch.apf.send_user_only && kvm_x86_ops.get_cpl(vcpu) == 0))
 		return false;
 
 	return true;
@@ -10499,7 +10543,10 @@ void kvm_arch_async_page_not_present(struct kvm_vcpu *vcpu,
 void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
 				 struct kvm_async_pf *work)
 {
-	struct x86_exception fault;
+	struct kvm_lapic_irq irq = {
+		.delivery_mode = APIC_DM_FIXED,
+		.vector = vcpu->arch.apf.vec
+	};
 
 	if (work->wakeup_all)
 		work->arch.token = ~0; /* broadcast wakeup */
@@ -10507,26 +10554,20 @@ void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
 		kvm_del_async_pf_gfn(vcpu, work->arch.gfn);
 	trace_kvm_async_pf_ready(work->arch.token, work->cr2_or_gpa);
 
-	if (vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED &&
-	    !apf_put_user_ready(vcpu, work->arch.token)) {
-			fault.vector = PF_VECTOR;
-			fault.error_code_valid = true;
-			fault.error_code = 0;
-			fault.nested_page_fault = false;
-			fault.address = work->arch.token;
-			fault.async_page_fault = true;
-			kvm_inject_page_fault(vcpu, &fault);
-	}
+	if (kvm_pv_async_pf_enabled(vcpu) &&
+	    !apf_put_user_ready(vcpu, work->arch.token))
+		kvm_apic_set_irq(vcpu, &irq, NULL);
+
 	vcpu->arch.apf.halted = false;
 	vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
 }
 
 bool kvm_arch_can_dequeue_async_page_present(struct kvm_vcpu *vcpu)
 {
-	if (!(vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED))
+	if (!kvm_pv_async_pf_enabled(vcpu))
 		return true;
 	else
-		return kvm_can_do_async_pf(vcpu);
+		return apf_pageready_slot_free(vcpu);
 }
 
 void kvm_arch_start_assignment(struct kvm *kvm)