kvm: x86: Defer setting of CR2 until #PF delivery

When exception payloads are enabled by userspace (which is not yet
possible) and a #PF is raised in L2, defer the setting of CR2 until
the #PF is delivered. This allows the L1 hypervisor to intercept the
fault before CR2 is modified.

For backwards compatibility, when exception payloads are not enabled
by userspace, kvm_multiple_exception modifies CR2 when the #PF
exception is raised.

Reported-by: Jim Mattson <jmattson@google.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

4 files changed, 62 insertions, 21 deletions

diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 47b07211c5b1..1086bebaeea3 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -805,6 +805,8 @@ static void svm_queue_exception(struct kvm_vcpu *vcpu)
 	    nested_svm_check_exception(svm, nr, has_error_code, error_code))
 		return;
 
+	kvm_deliver_exception_payload(&svm->vcpu);
+
 	if (nr == BP_VECTOR && !static_cpu_has(X86_FEATURE_NRIPS)) {
 		unsigned long rip, old_rip = kvm_rip_read(&svm->vcpu);
 
@@ -2965,16 +2967,13 @@ static int nested_svm_check_exception(struct vcpu_svm *svm, unsigned nr,
 	svm->vmcb->control.exit_info_1 = error_code;
 
 	/*
-	 * FIXME: we should not write CR2 when L1 intercepts an L2 #PF exception.
-	 * The fix is to add the ancillary datum (CR2 or DR6) to structs
-	 * kvm_queued_exception and kvm_vcpu_events, so that CR2 and DR6 can be
-	 * written only when inject_pending_event runs (DR6 would written here
-	 * too).  This should be conditional on a new capability---if the
-	 * capability is disabled, kvm_multiple_exception would write the
-	 * ancillary information to CR2 or DR6, for backwards ABI-compatibility.
+	 * EXITINFO2 is undefined for all exception intercepts other
+	 * than #PF.
 	 */
 	if (svm->vcpu.arch.exception.nested_apf)
 		svm->vmcb->control.exit_info_2 = svm->vcpu.arch.apf.nested_apf_token;
+	else if (svm->vcpu.arch.exception.has_payload)
+		svm->vmcb->control.exit_info_2 = svm->vcpu.arch.exception.payload;
 	else
 		svm->vmcb->control.exit_info_2 = svm->vcpu.arch.cr2;
 
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 8d7c60faaacd..0c7efb1de29d 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -3292,27 +3292,24 @@ static int nested_vmx_check_exception(struct kvm_vcpu *vcpu, unsigned long *exit
 {
 	struct vmcs12 *vmcs12 = get_vmcs12(vcpu);
 	unsigned int nr = vcpu->arch.exception.nr;
+	bool has_payload = vcpu->arch.exception.has_payload;
+	unsigned long payload = vcpu->arch.exception.payload;
 
 	if (nr == PF_VECTOR) {
 		if (vcpu->arch.exception.nested_apf) {
 			*exit_qual = vcpu->arch.apf.nested_apf_token;
 			return 1;
 		}
-		/*
-		 * FIXME: we must not write CR2 when L1 intercepts an L2 #PF exception.
-		 * The fix is to add the ancillary datum (CR2 or DR6) to structs
-		 * kvm_queued_exception and kvm_vcpu_events, so that CR2 and DR6
-		 * can be written only when inject_pending_event runs.  This should be
-		 * conditional on a new capability---if the capability is disabled,
-		 * kvm_multiple_exception would write the ancillary information to
-		 * CR2 or DR6, for backwards ABI-compatibility.
-		 */
 		if (nested_vmx_is_page_fault_vmexit(vmcs12,
 						    vcpu->arch.exception.error_code)) {
-			*exit_qual = vcpu->arch.cr2;
+			*exit_qual = has_payload ? payload : vcpu->arch.cr2;
 			return 1;
 		}
 	} else {
+		/*
+		 * FIXME: we must not write DR6 when L1 intercepts an
+		 * L2 #DB exception.
+		 */
 		if (vmcs12->exception_bitmap & (1u << nr)) {
 			if (nr == DB_VECTOR) {
 				*exit_qual = vcpu->arch.dr6;
@@ -3349,6 +3346,8 @@ static void vmx_queue_exception(struct kvm_vcpu *vcpu)
 	u32 error_code = vcpu->arch.exception.error_code;
 	u32 intr_info = nr | INTR_INFO_VALID_MASK;
 
+	kvm_deliver_exception_payload(vcpu);
+
 	if (has_error_code) {
 		vmcs_write32(VM_ENTRY_EXCEPTION_ERROR_CODE, error_code);
 		intr_info |= INTR_INFO_DELIVER_CODE_MASK;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 25a2bac920bb..4e9536c335d6 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -400,6 +400,26 @@ static int exception_type(int vector)
 	return EXCPT_FAULT;
 }
 
+void kvm_deliver_exception_payload(struct kvm_vcpu *vcpu)
+{
+	unsigned nr = vcpu->arch.exception.nr;
+	bool has_payload = vcpu->arch.exception.has_payload;
+	unsigned long payload = vcpu->arch.exception.payload;
+
+	if (!has_payload)
+		return;
+
+	switch (nr) {
+	case PF_VECTOR:
+		vcpu->arch.cr2 = payload;
+		break;
+	}
+
+	vcpu->arch.exception.has_payload = false;
+	vcpu->arch.exception.payload = 0;
+}
+EXPORT_SYMBOL_GPL(kvm_deliver_exception_payload);
+
 static void kvm_multiple_exception(struct kvm_vcpu *vcpu,
 		unsigned nr, bool has_error, u32 error_code,
 	        bool has_payload, unsigned long payload, bool reinject)
@@ -441,6 +461,18 @@ static void kvm_multiple_exception(struct kvm_vcpu *vcpu,
 		vcpu->arch.exception.error_code = error_code;
 		vcpu->arch.exception.has_payload = has_payload;
 		vcpu->arch.exception.payload = payload;
+		/*
+		 * In guest mode, payload delivery should be deferred,
+		 * so that the L1 hypervisor can intercept #PF before
+		 * CR2 is modified.  However, for ABI compatibility
+		 * with KVM_GET_VCPU_EVENTS and KVM_SET_VCPU_EVENTS,
+		 * we can't delay payload delivery unless userspace
+		 * has enabled this functionality via the per-VM
+		 * capability, KVM_CAP_EXCEPTION_PAYLOAD.
+		 */
+		if (!vcpu->kvm->arch.exception_payload_enabled ||
+		    !is_guest_mode(vcpu))
+			kvm_deliver_exception_payload(vcpu);
 		return;
 	}
 
@@ -486,6 +518,13 @@ void kvm_requeue_exception(struct kvm_vcpu *vcpu, unsigned nr)
 }
 EXPORT_SYMBOL_GPL(kvm_requeue_exception);
 
+static void kvm_queue_exception_e_p(struct kvm_vcpu *vcpu, unsigned nr,
+				    u32 error_code, unsigned long payload)
+{
+	kvm_multiple_exception(vcpu, nr, true, error_code,
+			       true, payload, false);
+}
+
 int kvm_complete_insn_gp(struct kvm_vcpu *vcpu, int err)
 {
 	if (err)
@@ -502,11 +541,13 @@ void kvm_inject_page_fault(struct kvm_vcpu *vcpu, struct x86_exception *fault)
 	++vcpu->stat.pf_guest;
 	vcpu->arch.exception.nested_apf =
 		is_guest_mode(vcpu) && fault->async_page_fault;
-	if (vcpu->arch.exception.nested_apf)
+	if (vcpu->arch.exception.nested_apf) {
 		vcpu->arch.apf.nested_apf_token = fault->address;
-	else
-		vcpu->arch.cr2 = fault->address;
-	kvm_queue_exception_e(vcpu, PF_VECTOR, fault->error_code);
+		kvm_queue_exception_e(vcpu, PF_VECTOR, fault->error_code);
+	} else {
+		kvm_queue_exception_e_p(vcpu, PF_VECTOR, fault->error_code,
+					fault->address);
+	}
 }
 EXPORT_SYMBOL_GPL(kvm_inject_page_fault);
 
diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
index 67b9568613f3..224cd0a47568 100644
--- a/arch/x86/kvm/x86.h
+++ b/arch/x86/kvm/x86.h
@@ -266,6 +266,8 @@ int kvm_write_guest_virt_system(struct kvm_vcpu *vcpu,
 
 int handle_ud(struct kvm_vcpu *vcpu);
 
+void kvm_deliver_exception_payload(struct kvm_vcpu *vcpu);
+
 void kvm_vcpu_mtrr_init(struct kvm_vcpu *vcpu);
 u8 kvm_mtrr_get_guest_memory_type(struct kvm_vcpu *vcpu, gfn_t gfn);
 bool kvm_mtrr_valid(struct kvm_vcpu *vcpu, u32 msr, u64 data);