KVM: nVMX: trace nested VM-Enter failures detected by H/W

Use the recently added tracepoint for logging nested VM-Enter failures
instead of spamming the kernel log when hardware detects a consistency
check failure.  Take the opportunity to print the name of the error code
instead of dumping the raw hex number, but limit the symbol table to
error codes that can reasonably be encountered by KVM.

Add an equivalent tracepoint in nested_vmx_check_vmentry_hw(), e.g. so
that tracing of "invalid control field" errors isn't suppressed when
nested early checks are enabled.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

3 files changed, 30 insertions, 8 deletions

diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
index a39136b0d509..b15e6465870f 100644
--- a/arch/x86/include/asm/vmx.h
+++ b/arch/x86/include/asm/vmx.h
@@ -562,6 +562,20 @@ enum vm_instruction_error_number {
 	VMXERR_INVALID_OPERAND_TO_INVEPT_INVVPID = 28,
 };
 
+/*
+ * VM-instruction errors that can be encountered on VM-Enter, used to trace
+ * nested VM-Enter failures reported by hardware.  Errors unique to VM-Enter
+ * from a SMI Transfer Monitor are not included as things have gone seriously
+ * sideways if we get one of those...
+ */
+#define VMX_VMENTER_INSTRUCTION_ERRORS \
+	{ VMXERR_VMLAUNCH_NONCLEAR_VMCS,		"VMLAUNCH_NONCLEAR_VMCS" }, \
+	{ VMXERR_VMRESUME_NONLAUNCHED_VMCS,		"VMRESUME_NONLAUNCHED_VMCS" }, \
+	{ VMXERR_VMRESUME_AFTER_VMXOFF,			"VMRESUME_AFTER_VMXOFF" }, \
+	{ VMXERR_ENTRY_INVALID_CONTROL_FIELD,		"VMENTRY_INVALID_CONTROL_FIELD" }, \
+	{ VMXERR_ENTRY_INVALID_HOST_STATE_FIELD,	"VMENTRY_INVALID_HOST_STATE_FIELD" }, \
+	{ VMXERR_ENTRY_EVENTS_BLOCKED_BY_MOV_SS,	"VMENTRY_EVENTS_BLOCKED_BY_MOV_SS" }
+
 enum vmx_l1d_flush_state {
 	VMENTER_L1D_FLUSH_AUTO,
 	VMENTER_L1D_FLUSH_NEVER,
diff --git a/arch/x86/kvm/trace.h b/arch/x86/kvm/trace.h
index 25ea92873e49..7c741a0c5f80 100644
--- a/arch/x86/kvm/trace.h
+++ b/arch/x86/kvm/trace.h
@@ -1479,18 +1479,21 @@ TRACE_EVENT(kvm_pv_tlb_flush,
  * Tracepoint for failed nested VMX VM-Enter.
  */
 TRACE_EVENT(kvm_nested_vmenter_failed,
-	TP_PROTO(const char *msg),
-	TP_ARGS(msg),
+	TP_PROTO(const char *msg, u32 err),
+	TP_ARGS(msg, err),
 
 	TP_STRUCT__entry(
 		__field(const char *, msg)
+		__field(u32, err)
 	),
 
 	TP_fast_assign(
 		__entry->msg = msg;
+		__entry->err = err;
 	),
 
-	TP_printk("%s", __entry->msg)
+	TP_printk("%s%s", __entry->msg, !__entry->err ? "" :
+		__print_symbolic(__entry->err, VMX_VMENTER_INSTRUCTION_ERRORS))
 );
 
 #endif /* _TRACE_KVM_H */
diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
index bdfc5570bab8..ad2453317c4b 100644
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -23,7 +23,7 @@ module_param(nested_early_check, bool, S_IRUGO);
 ({									\
 	bool failed = (consistency_check);				\
 	if (failed)							\
-		trace_kvm_nested_vmenter_failed(#consistency_check);	\
+		trace_kvm_nested_vmenter_failed(#consistency_check, 0);	\
 	failed;								\
 })
 
@@ -2845,9 +2845,13 @@ static int nested_vmx_check_vmentry_hw(struct kvm_vcpu *vcpu)
 		vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr);
 
 	if (vm_fail) {
+		u32 error = vmcs_read32(VM_INSTRUCTION_ERROR);
+
 		preempt_enable();
-		WARN_ON_ONCE(vmcs_read32(VM_INSTRUCTION_ERROR) !=
-			     VMXERR_ENTRY_INVALID_CONTROL_FIELD);
+
+		trace_kvm_nested_vmenter_failed(
+			"early hardware check VM-instruction error: ", error);
+		WARN_ON_ONCE(error != VMXERR_ENTRY_INVALID_CONTROL_FIELD);
 		return 1;
 	}
 
@@ -5259,8 +5263,9 @@ bool nested_vmx_exit_reflected(struct kvm_vcpu *vcpu, u32 exit_reason)
 		return false;
 
 	if (unlikely(vmx->fail)) {
-		pr_info_ratelimited("%s failed vm entry %x\n", __func__,
-				    vmcs_read32(VM_INSTRUCTION_ERROR));
+		trace_kvm_nested_vmenter_failed(
+			"hardware VM-instruction error: ",
+			vmcs_read32(VM_INSTRUCTION_ERROR));
 		return true;
 	}