Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1515523 > unrolled thread

[PATCH v2] KVM: nVMX: add tracepoint for vmwrite

Back to article view | Back to linux.kernel

Contents

  [PATCH v2] KVM: nVMX: add tracepoint for vmwrite Paolo Bonzini <pbonzini@redhat.com> - 2016-11-05 09:50 +0100
    Re: [PATCH v2] KVM: nVMX: add tracepoint for vmwrite Ladi Prosek <lprosek@redhat.com> - 2016-11-07 15:00 +0100

#1515523 — [PATCH v2] KVM: nVMX: add tracepoint for vmwrite

Both the field and the value are in registers, so they are not clear from
existing tracepoints.

Cc: Ladi Prosek <lprosek@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/kvm/trace.h | 23 +++++++++++++++++++++++
 arch/x86/kvm/vmx.c   |  2 ++
 arch/x86/kvm/x86.c   |  1 +
 3 files changed, 26 insertions(+)

diff --git a/arch/x86/kvm/trace.h b/arch/x86/kvm/trace.h
index 0a6cc6754ec5..b3d4dec78ae2 100644
--- a/arch/x86/kvm/trace.h
+++ b/arch/x86/kvm/trace.h
@@ -723,6 +723,29 @@ TRACE_EVENT(kvm_skinit,
 	flags;						\
 	})
 
+/*
+ * Tracepoint for nested #vmexit because of vmwrite
+ */
+TRACE_EVENT(kvm_vmwrite,
+	    TP_PROTO(__u64 vmptr, __u16 field, __u64 value),
+	    TP_ARGS(vmptr, field, value),
+
+	TP_STRUCT__entry(
+		__field(	__u64,	vmptr	)
+		__field(	__u16,	field	)
+		__field(	__u64,	value	)
+	),
+
+	TP_fast_assign(
+		__entry->vmptr		=	vmptr;
+		__entry->field		=	field;
+		__entry->value		=	value;
+	),
+
+	TP_printk("vmcs: 0x%016llx field: 0x%04x value: 0x%llx",
+		  __entry->vmptr, __entry->field, __entry->value)
+);
+
 TRACE_EVENT(kvm_emulate_insn,
 	TP_PROTO(struct kvm_vcpu *vcpu, __u8 failed),
 	TP_ARGS(vcpu, failed),
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5382b82462fc..e7d01d7d350b 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -7474,6 +7474,7 @@ static int handle_vmwrite(struct kvm_vcpu *vcpu)
 {
 	unsigned long field;
 	gva_t gva;
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	unsigned long exit_qualification = vmcs_readl(EXIT_QUALIFICATION);
 	u32 vmx_instruction_info = vmcs_read32(VMX_INSTRUCTION_INFO);
 	/* The value to write might be 32 or 64 bits, depending on L1's long
@@ -7512,6 +7513,7 @@ static int handle_vmwrite(struct kvm_vcpu *vcpu)
 		return 1;
 	}
 
+	trace_kvm_vmwrite(vmx->nested.current_vmptr, field, field_value);
 	if (vmcs12_write_any(vcpu, field, field_value) < 0) {
 		nested_vmx_failValid(vcpu, VMXERR_UNSUPPORTED_VMCS_COMPONENT);
 		skip_emulated_instruction(vcpu);
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 7e30c720d0c5..2be66e90f297 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -8495,3 +8495,4 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_pml_full);
 EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_pi_irte_update);
 EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_avic_unaccelerated_access);
 EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_avic_incomplete_ipi);
+EXPORT_TRACEPOINT_SYMBOL_GPL(kvm_vmwrite);
-- 
2.7.4

[toc] | [next] | [standalone]

#1516214

On Sat, Nov 5, 2016 at 9:44 AM, Paolo Bonzini <pbonzini@redhat.com> wrote:
> Both the field and the value are in registers, so they are not clear from
> existing tracepoints.
>
> Cc: Ladi Prosek <lprosek@redhat.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  arch/x86/kvm/trace.h | 23 +++++++++++++++++++++++
>  arch/x86/kvm/vmx.c   |  2 ++
>  arch/x86/kvm/x86.c   |  1 +
>  3 files changed, 26 insertions(+)

Works for me, thanks!

Tested-by: Ladi Prosek <lprosek@redhat.com>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web