[PATCH v2 0/6] KVM: nVMX: Enlightened VMCS for Hyper-V on KVM

[PATCH v2 0/6] KVM: nVMX: Enlightened VMCS for Hyper-V on KVM

[Vitaly Kuznetsov]

Changes since v1 [Liran Alon]:
- Split PATCH2 into two.
- Add a comment explaining eVMCS versioning.
- Convert evmcs_needs_write() to inline function.
- Drop 'unlikely' in vmx_vcpu_run().
- Add Reviewed-by tags.

Description:

This is an initial implementation of Enlightened VMCS for nested Hyper-V on
KVM. Using it helps to spare 1500 cpu cycles for nested vmexit (tight cpuid
loop in WS2016 with Hyper-V role on KVM: 15200 cycles -> 13700 cycles).

Ladi Prosek (1):
  KVM: hyperv: define VP assist page helpers

Vitaly Kuznetsov (5):
  KVM: VMX: refactor evmcs_sanitize_exec_ctrls()
  KVM: nVMX: add KVM_CAP_HYPERV_ENLIGHTENED_VMCS capability
  KVM: nVMX: add enlightened VMCS state
  KVM: nVMX: implement enlightened VMPTRLD and VMCLEAR
  KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

 arch/x86/include/asm/kvm_host.h |   3 +
 arch/x86/kvm/hyperv.c           |  23 +-
 arch/x86/kvm/hyperv.h           |   4 +
 arch/x86/kvm/lapic.c            |   4 +-
 arch/x86/kvm/lapic.h            |   2 +-
 arch/x86/kvm/svm.c              |   9 +
 arch/x86/kvm/vmx.c              | 830 +++++++++++++++++++++++++++++++++-------
 arch/x86/kvm/x86.c              |  17 +-
 include/uapi/linux/kvm.h        |   1 +
 9 files changed, 742 insertions(+), 151 deletions(-)

-- 
2.14.4

[PATCH v2 1/6] KVM: hyperv: define VP assist page helpers

[Vitaly Kuznetsov]

From: Ladi Prosek <lprosek@redhat.com>

The state related to the VP assist page is still managed by the LAPIC
code in the pv_eoi field.

Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Liran Alon <liran.alon@oracle.com>
---
 arch/x86/kvm/hyperv.c | 23 +++++++++++++++++++++--
 arch/x86/kvm/hyperv.h |  4 ++++
 arch/x86/kvm/lapic.c  |  4 ++--
 arch/x86/kvm/lapic.h  |  2 +-
 arch/x86/kvm/x86.c    |  2 +-
 5 files changed, 29 insertions(+), 6 deletions(-)

diff --git a/arch/x86/kvm/hyperv.c b/arch/x86/kvm/hyperv.c
index af8caf965baa..af7851077925 100644
--- a/arch/x86/kvm/hyperv.c
+++ b/arch/x86/kvm/hyperv.c
@@ -688,6 +688,24 @@ void kvm_hv_vcpu_uninit(struct kvm_vcpu *vcpu)
 		stimer_cleanup(&hv_vcpu->stimer[i]);
 }
 
+bool kvm_hv_assist_page_enabled(struct kvm_vcpu *vcpu)
+{
+	if (!(vcpu->arch.hyperv.hv_vapic & HV_X64_MSR_VP_ASSIST_PAGE_ENABLE))
+		return false;
+	return vcpu->arch.pv_eoi.msr_val & KVM_MSR_ENABLED;
+}
+EXPORT_SYMBOL_GPL(kvm_hv_assist_page_enabled);
+
+bool kvm_hv_get_assist_page(struct kvm_vcpu *vcpu,
+			    struct hv_vp_assist_page *assist_page)
+{
+	if (!kvm_hv_assist_page_enabled(vcpu))
+		return false;
+	return !kvm_read_guest_cached(vcpu->kvm, &vcpu->arch.pv_eoi.data,
+				      assist_page, sizeof(*assist_page));
+}
+EXPORT_SYMBOL_GPL(kvm_hv_get_assist_page);
+
 static void stimer_prepare_msg(struct kvm_vcpu_hv_stimer *stimer)
 {
 	struct hv_message *msg = &stimer->msg;
@@ -1048,7 +1066,7 @@ static int kvm_hv_set_msr(struct kvm_vcpu *vcpu, u32 msr, u64 data, bool host)
 
 		if (!(data & HV_X64_MSR_VP_ASSIST_PAGE_ENABLE)) {
 			hv->hv_vapic = data;
-			if (kvm_lapic_enable_pv_eoi(vcpu, 0))
+			if (kvm_lapic_enable_pv_eoi(vcpu, 0, 0))
 				return 1;
 			break;
 		}
@@ -1061,7 +1079,8 @@ static int kvm_hv_set_msr(struct kvm_vcpu *vcpu, u32 msr, u64 data, bool host)
 		hv->hv_vapic = data;
 		kvm_vcpu_mark_page_dirty(vcpu, gfn);
 		if (kvm_lapic_enable_pv_eoi(vcpu,
-					    gfn_to_gpa(gfn) | KVM_MSR_ENABLED))
+					    gfn_to_gpa(gfn) | KVM_MSR_ENABLED,
+					    sizeof(struct hv_vp_assist_page)))
 			return 1;
 		break;
 	}
diff --git a/arch/x86/kvm/hyperv.h b/arch/x86/kvm/hyperv.h
index 837465d69c6d..db825bb7efc7 100644
--- a/arch/x86/kvm/hyperv.h
+++ b/arch/x86/kvm/hyperv.h
@@ -62,6 +62,10 @@ void kvm_hv_vcpu_init(struct kvm_vcpu *vcpu);
 void kvm_hv_vcpu_postcreate(struct kvm_vcpu *vcpu);
 void kvm_hv_vcpu_uninit(struct kvm_vcpu *vcpu);
 
+bool kvm_hv_assist_page_enabled(struct kvm_vcpu *vcpu);
+bool kvm_hv_get_assist_page(struct kvm_vcpu *vcpu,
+			    struct hv_vp_assist_page *assist_page);
+
 static inline struct kvm_vcpu_hv_stimer *vcpu_to_stimer(struct kvm_vcpu *vcpu,
 							int timer_index)
 {
diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index b5cd8465d44f..1ee7e0a68f9f 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -2552,7 +2552,7 @@ int kvm_hv_vapic_msr_read(struct kvm_vcpu *vcpu, u32 reg, u64 *data)
 	return 0;
 }
 
-int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data)
+int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len)
 {
 	u64 addr = data & ~KVM_MSR_ENABLED;
 	if (!IS_ALIGNED(addr, 4))
@@ -2562,7 +2562,7 @@ int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data)
 	if (!pv_eoi_enabled(vcpu))
 		return 0;
 	return kvm_gfn_to_hva_cache_init(vcpu->kvm, &vcpu->arch.pv_eoi.data,
-					 addr, sizeof(u8));
+					 addr, len);
 }
 
 void kvm_apic_accept_events(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/lapic.h b/arch/x86/kvm/lapic.h
index ed0ed39abd36..ff6ef9c3d760 100644
--- a/arch/x86/kvm/lapic.h
+++ b/arch/x86/kvm/lapic.h
@@ -120,7 +120,7 @@ static inline bool kvm_hv_vapic_assist_page_enabled(struct kvm_vcpu *vcpu)
 	return vcpu->arch.hyperv.hv_vapic & HV_X64_MSR_VP_ASSIST_PAGE_ENABLE;
 }
 
-int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data);
+int kvm_lapic_enable_pv_eoi(struct kvm_vcpu *vcpu, u64 data, unsigned long len);
 void kvm_lapic_init(void);
 void kvm_lapic_exit(void);
 
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 0046aa70205a..c95953e61e3e 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2443,7 +2443,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 
 		break;
 	case MSR_KVM_PV_EOI_EN:
-		if (kvm_lapic_enable_pv_eoi(vcpu, data))
+		if (kvm_lapic_enable_pv_eoi(vcpu, data, sizeof(u8)))
 			return 1;
 		break;
 
-- 
2.14.4

[PATCH v2 2/6] KVM: VMX: refactor evmcs_sanitize_exec_ctrls()

[Vitaly Kuznetsov]

Split off EVMCS1_UNSUPPORTED_* macros so we can re-use them when
enabling Enlightened VMCS for Hyper-V on KVM.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
 arch/x86/kvm/vmx.c | 108 +++++++++++++++++++++++------------------------------
 1 file changed, 47 insertions(+), 61 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 559a12b6184d..2bde41549bff 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1186,6 +1186,49 @@ DEFINE_STATIC_KEY_FALSE(enable_evmcs);
 
 #define KVM_EVMCS_VERSION 1
 
+/*
+ * Enlightened VMCSv1 doesn't support these:
+ *
+ *	POSTED_INTR_NV                  = 0x00000002,
+ *	GUEST_INTR_STATUS               = 0x00000810,
+ *	APIC_ACCESS_ADDR		= 0x00002014,
+ *	POSTED_INTR_DESC_ADDR           = 0x00002016,
+ *	EOI_EXIT_BITMAP0                = 0x0000201c,
+ *	EOI_EXIT_BITMAP1                = 0x0000201e,
+ *	EOI_EXIT_BITMAP2                = 0x00002020,
+ *	EOI_EXIT_BITMAP3                = 0x00002022,
+ *	GUEST_PML_INDEX			= 0x00000812,
+ *	PML_ADDRESS			= 0x0000200e,
+ *	VM_FUNCTION_CONTROL             = 0x00002018,
+ *	EPTP_LIST_ADDRESS               = 0x00002024,
+ *	VMREAD_BITMAP                   = 0x00002026,
+ *	VMWRITE_BITMAP                  = 0x00002028,
+ *
+ *	TSC_MULTIPLIER                  = 0x00002032,
+ *	PLE_GAP                         = 0x00004020,
+ *	PLE_WINDOW                      = 0x00004022,
+ *	VMX_PREEMPTION_TIMER_VALUE      = 0x0000482E,
+ *      GUEST_IA32_PERF_GLOBAL_CTRL     = 0x00002808,
+ *      HOST_IA32_PERF_GLOBAL_CTRL      = 0x00002c04,
+ *
+ * Currently unsupported in KVM:
+ *	GUEST_IA32_RTIT_CTL		= 0x00002814,
+ */
+#define EVMCS1_UNSUPPORTED_PINCTRL (PIN_BASED_POSTED_INTR | \
+				    PIN_BASED_VMX_PREEMPTION_TIMER)
+#define EVMCS1_UNSUPPORTED_2NDEXEC					\
+	(SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |				\
+	 SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |			\
+	 SECONDARY_EXEC_APIC_REGISTER_VIRT |				\
+	 SECONDARY_EXEC_ENABLE_PML |					\
+	 SECONDARY_EXEC_ENABLE_VMFUNC |					\
+	 SECONDARY_EXEC_SHADOW_VMCS |					\
+	 SECONDARY_EXEC_TSC_SCALING |					\
+	 SECONDARY_EXEC_PAUSE_LOOP_EXITING)
+#define EVMCS1_UNSUPPORTED_VMEXIT_CTRL (VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL)
+#define EVMCS1_UNSUPPORTED_VMENTRY_CTRL (VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL)
+#define EVMCS1_UNSUPPORTED_VMFUNC (VMX_VMFUNC_EPTP_SWITCHING)
+
 #if IS_ENABLED(CONFIG_HYPERV)
 static bool __read_mostly enlightened_vmcs = true;
 module_param(enlightened_vmcs, bool, 0444);
@@ -1278,69 +1321,12 @@ static void evmcs_load(u64 phys_addr)
 
 static void evmcs_sanitize_exec_ctrls(struct vmcs_config *vmcs_conf)
 {
-	/*
-	 * Enlightened VMCSv1 doesn't support these:
-	 *
-	 *	POSTED_INTR_NV                  = 0x00000002,
-	 *	GUEST_INTR_STATUS               = 0x00000810,
-	 *	APIC_ACCESS_ADDR		= 0x00002014,
-	 *	POSTED_INTR_DESC_ADDR           = 0x00002016,
-	 *	EOI_EXIT_BITMAP0                = 0x0000201c,
-	 *	EOI_EXIT_BITMAP1                = 0x0000201e,
-	 *	EOI_EXIT_BITMAP2                = 0x00002020,
-	 *	EOI_EXIT_BITMAP3                = 0x00002022,
-	 */
-	vmcs_conf->pin_based_exec_ctrl &= ~PIN_BASED_POSTED_INTR;
-	vmcs_conf->cpu_based_2nd_exec_ctrl &=
-		~SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY;
-	vmcs_conf->cpu_based_2nd_exec_ctrl &=
-		~SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES;
-	vmcs_conf->cpu_based_2nd_exec_ctrl &=
-		~SECONDARY_EXEC_APIC_REGISTER_VIRT;
-
-	/*
-	 *	GUEST_PML_INDEX			= 0x00000812,
-	 *	PML_ADDRESS			= 0x0000200e,
-	 */
-	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~SECONDARY_EXEC_ENABLE_PML;
+	vmcs_conf->pin_based_exec_ctrl &= ~EVMCS1_UNSUPPORTED_PINCTRL;
+	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~EVMCS1_UNSUPPORTED_2NDEXEC;
 
-	/*	VM_FUNCTION_CONTROL             = 0x00002018, */
-	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~SECONDARY_EXEC_ENABLE_VMFUNC;
-
-	/*
-	 *	EPTP_LIST_ADDRESS               = 0x00002024,
-	 *	VMREAD_BITMAP                   = 0x00002026,
-	 *	VMWRITE_BITMAP                  = 0x00002028,
-	 */
-	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~SECONDARY_EXEC_SHADOW_VMCS;
+	vmcs_conf->vmexit_ctrl &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL;
+	vmcs_conf->vmentry_ctrl &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL;
 
-	/*
-	 *	TSC_MULTIPLIER                  = 0x00002032,
-	 */
-	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~SECONDARY_EXEC_TSC_SCALING;
-
-	/*
-	 *	PLE_GAP                         = 0x00004020,
-	 *	PLE_WINDOW                      = 0x00004022,
-	 */
-	vmcs_conf->cpu_based_2nd_exec_ctrl &= ~SECONDARY_EXEC_PAUSE_LOOP_EXITING;
-
-	/*
-	 *	VMX_PREEMPTION_TIMER_VALUE      = 0x0000482E,
-	 */
-	vmcs_conf->pin_based_exec_ctrl &= ~PIN_BASED_VMX_PREEMPTION_TIMER;
-
-	/*
-	 *      GUEST_IA32_PERF_GLOBAL_CTRL     = 0x00002808,
-	 *      HOST_IA32_PERF_GLOBAL_CTRL      = 0x00002c04,
-	 */
-	vmcs_conf->vmexit_ctrl &= ~VM_EXIT_LOAD_IA32_PERF_GLOBAL_CTRL;
-	vmcs_conf->vmentry_ctrl &= ~VM_ENTRY_LOAD_IA32_PERF_GLOBAL_CTRL;
-
-	/*
-	 * Currently unsupported in KVM:
-	 *	GUEST_IA32_RTIT_CTL		= 0x00002814,
-	 */
 }
 #else /* !IS_ENABLED(CONFIG_HYPERV) */
 static inline void evmcs_write64(unsigned long field, u64 value) {}
-- 
2.14.4

[PATCH v2 3/6] KVM: nVMX: add KVM_CAP_HYPERV_ENLIGHTENED_VMCS capability

[Vitaly Kuznetsov]

Enlightened VMCS is opt-in. The current version does not contain all
fields supported by nested VMX so we must not advertise the
corresponding VMX features if enlightened VMCS is enabled.

Userspace is given the enlightened VMCS version supported by KVM as
part of enabling KVM_CAP_HYPERV_ENLIGHTENED_VMCS. The version is to
be advertised to the nested hypervisor, currently done via a cpuid
leaf for Hyper-V.

Suggested-by: Ladi Prosek <lprosek@redhat.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Liran Alon <liran.alon@oracle.com>
---
Changes since v1:
- Add a comment explaining "<< 8) | 1" [Liran Alon]
---
 arch/x86/include/asm/kvm_host.h |  3 +++
 arch/x86/kvm/svm.c              |  9 +++++++++
 arch/x86/kvm/vmx.c              | 37 +++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c              | 15 +++++++++++++++
 include/uapi/linux/kvm.h        |  1 +
 5 files changed, 65 insertions(+)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index c13cd28d9d1b..f7f83fc9f3af 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1095,6 +1095,9 @@ struct kvm_x86_ops {
 	int (*mem_enc_unreg_region)(struct kvm *kvm, struct kvm_enc_region *argp);
 
 	int (*get_msr_feature)(struct kvm_msr_entry *entry);
+
+	int (*nested_enable_evmcs)(struct kvm_vcpu *vcpu,
+				   uint16_t *vmcs_version);
 };
 
 struct kvm_arch_async_pf {
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index f059a73f0fd0..2e8352747288 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -7035,6 +7035,13 @@ static int svm_unregister_enc_region(struct kvm *kvm,
 	return ret;
 }
 
+static int nested_enable_evmcs(struct kvm_vcpu *vcpu,
+				   uint16_t *vmcs_version)
+{
+	/* Intel-only feature */
+	return -ENODEV;
+}
+
 static struct kvm_x86_ops svm_x86_ops __ro_after_init = {
 	.cpu_has_kvm_support = has_svm,
 	.disabled_by_bios = is_disabled,
@@ -7161,6 +7168,8 @@ static struct kvm_x86_ops svm_x86_ops __ro_after_init = {
 	.mem_enc_op = svm_mem_enc_op,
 	.mem_enc_reg_region = svm_register_enc_region,
 	.mem_enc_unreg_region = svm_unregister_enc_region,
+
+	.nested_enable_evmcs = nested_enable_evmcs,
 };
 
 static int __init svm_init(void)
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 2bde41549bff..734de9e284c6 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -648,6 +648,13 @@ struct nested_vmx {
 
 	bool change_vmcs01_virtual_apic_mode;
 
+	/*
+	 * Enlightened VMCS has been enabled. It does not mean that L1 has to
+	 * use it. However, VMX features available to L1 will be limited based
+	 * on what the enlightened VMCS supports.
+	 */
+	bool enlightened_vmcs_enabled;
+
 	/* L2 must run next, and mustn't decide to exit to L1. */
 	bool nested_run_pending;
 
@@ -1340,6 +1347,34 @@ static inline void evmcs_sanitize_exec_ctrls(struct vmcs_config *vmcs_conf) {}
 static inline void evmcs_touch_msr_bitmap(void) {}
 #endif /* IS_ENABLED(CONFIG_HYPERV) */
 
+static int nested_enable_evmcs(struct kvm_vcpu *vcpu,
+			       uint16_t *vmcs_version)
+{
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
+
+	/* We don't support disabling the feature for simplicity. */
+	if (vmx->nested.enlightened_vmcs_enabled)
+		return 0;
+
+	vmx->nested.enlightened_vmcs_enabled = true;
+
+	/*
+	 * vmcs_version represents the range of supported Enlightened VMCS
+	 * versions: lower 8 bits is the minimal version, higher 8 bits is the
+	 * maximum supported version. KVM supports versions from 1 to
+	 * KVM_EVMCS_VERSION.
+	 */
+	*vmcs_version = (KVM_EVMCS_VERSION << 8) | 1;
+
+	vmx->nested.msrs.pinbased_ctls_high &= ~EVMCS1_UNSUPPORTED_PINCTRL;
+	vmx->nested.msrs.entry_ctls_high &= ~EVMCS1_UNSUPPORTED_VMENTRY_CTRL;
+	vmx->nested.msrs.exit_ctls_high &= ~EVMCS1_UNSUPPORTED_VMEXIT_CTRL;
+	vmx->nested.msrs.secondary_ctls_high &= ~EVMCS1_UNSUPPORTED_2NDEXEC;
+	vmx->nested.msrs.vmfunc_controls &= ~EVMCS1_UNSUPPORTED_VMFUNC;
+
+	return 0;
+}
+
 static inline bool is_exception_n(u32 intr_info, u8 vector)
 {
 	return (intr_info & (INTR_INFO_INTR_TYPE_MASK | INTR_INFO_VECTOR_MASK |
@@ -13033,6 +13068,8 @@ static struct kvm_x86_ops vmx_x86_ops __ro_after_init = {
 	.pre_enter_smm = vmx_pre_enter_smm,
 	.pre_leave_smm = vmx_pre_leave_smm,
 	.enable_smi_window = enable_smi_window,
+
+	.nested_enable_evmcs = nested_enable_evmcs,
 };
 
 static int __init vmx_init(void)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index c95953e61e3e..fa1aca7bfba1 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2874,6 +2874,7 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
 	case KVM_CAP_HYPERV_VP_INDEX:
 	case KVM_CAP_HYPERV_EVENTFD:
 	case KVM_CAP_HYPERV_TLBFLUSH:
+	case KVM_CAP_HYPERV_ENLIGHTENED_VMCS:
 	case KVM_CAP_PCI_SEGMENT:
 	case KVM_CAP_DEBUGREGS:
 	case KVM_CAP_X86_ROBUST_SINGLESTEP:
@@ -3651,6 +3652,10 @@ static int kvm_set_guest_paused(struct kvm_vcpu *vcpu)
 static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu,
 				     struct kvm_enable_cap *cap)
 {
+	int r;
+	uint16_t vmcs_version;
+	void __user *user_ptr;
+
 	if (cap->flags)
 		return -EINVAL;
 
@@ -3663,6 +3668,16 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu,
 			return -EINVAL;
 		return kvm_hv_activate_synic(vcpu, cap->cap ==
 					     KVM_CAP_HYPERV_SYNIC2);
+	case KVM_CAP_HYPERV_ENLIGHTENED_VMCS:
+		r = kvm_x86_ops->nested_enable_evmcs(vcpu, &vmcs_version);
+		if (!r) {
+			user_ptr = (void __user *)(uintptr_t)cap->args[0];
+			if (copy_to_user(user_ptr, &vmcs_version,
+					 sizeof(vmcs_version)))
+				r = -EFAULT;
+		}
+		return r;
+
 	default:
 		return -EINVAL;
 	}
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index b6270a3b38e9..5c4b79c1af19 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -949,6 +949,7 @@ struct kvm_ppc_resize_hpt {
 #define KVM_CAP_GET_MSR_FEATURES 153
 #define KVM_CAP_HYPERV_EVENTFD 154
 #define KVM_CAP_HYPERV_TLBFLUSH 155
+#define KVM_CAP_HYPERV_ENLIGHTENED_VMCS 156
 
 #ifdef KVM_CAP_IRQ_ROUTING
 
-- 
2.14.4

[PATCH v2 4/6] KVM: nVMX: add enlightened VMCS state

[Vitaly Kuznetsov]

Adds hv_evmcs pointer and implement copy_enlightened_to_vmcs12() and
copy_enlightened_to_vmcs12().

prepare_vmcs02()/prepare_vmcs02_full() separation is not valid for
Enlightened VMCS, do full sync for now.

Suggested-by: Ladi Prosek <lprosek@redhat.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
Changes since v1:
- Drop 'unlikely' [Liran Alon]
---
 arch/x86/kvm/vmx.c | 431 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 417 insertions(+), 14 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 734de9e284c6..47c34c2c3ba7 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -640,10 +640,10 @@ struct nested_vmx {
 	 */
 	struct vmcs12 *cached_vmcs12;
 	/*
-	 * Indicates if the shadow vmcs must be updated with the
-	 * data hold by vmcs12
+	 * Indicates if the shadow vmcs or enlightened vmcs must be updated
+	 * with the data held by struct vmcs12.
 	 */
-	bool sync_shadow_vmcs;
+	bool need_vmcs12_sync;
 	bool dirty_vmcs12;
 
 	bool change_vmcs01_virtual_apic_mode;
@@ -689,6 +689,8 @@ struct nested_vmx {
 		/* in guest mode on SMM entry? */
 		bool guest_mode;
 	} smm;
+
+	struct hv_enlightened_vmcs *hv_evmcs;
 };
 
 #define POSTED_INTR_ON  0
@@ -8015,7 +8017,7 @@ static inline void nested_release_vmcs12(struct vcpu_vmx *vmx)
 		/* copy to memory all shadowed fields in case
 		   they were modified */
 		copy_shadow_to_vmcs12(vmx);
-		vmx->nested.sync_shadow_vmcs = false;
+		vmx->nested.need_vmcs12_sync = false;
 		vmx_disable_shadow_vmcs(vmx);
 	}
 	vmx->nested.posted_intr_nv = -1;
@@ -8192,6 +8194,393 @@ static inline int vmcs12_write_any(struct kvm_vcpu *vcpu,
 
 }
 
+static int copy_enlightened_to_vmcs12(struct vcpu_vmx *vmx, bool full)
+{
+	struct vmcs12 *vmcs12 = vmx->nested.cached_vmcs12;
+	struct hv_enlightened_vmcs *evmcs = vmx->nested.hv_evmcs;
+
+	/* HV_VMX_ENLIGHTENED_CLEAN_FIELD_NONE */
+	vmcs12->tpr_threshold = evmcs->tpr_threshold;
+	vmcs12->guest_rip = evmcs->guest_rip;
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_BASIC))) {
+		vmcs12->guest_rsp = evmcs->guest_rsp;
+		vmcs12->guest_rflags = evmcs->guest_rflags;
+		vmcs12->guest_interruptibility_info =
+			evmcs->guest_interruptibility_info;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_PROC))) {
+		vmcs12->cpu_based_vm_exec_control =
+			evmcs->cpu_based_vm_exec_control;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_PROC))) {
+		vmcs12->exception_bitmap = evmcs->exception_bitmap;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_ENTRY))) {
+		vmcs12->vm_entry_controls = evmcs->vm_entry_controls;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_EVENT))) {
+		vmcs12->vm_entry_intr_info_field =
+			evmcs->vm_entry_intr_info_field;
+		vmcs12->vm_entry_exception_error_code =
+			evmcs->vm_entry_exception_error_code;
+		vmcs12->vm_entry_instruction_len =
+			evmcs->vm_entry_instruction_len;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+		      HV_VMX_ENLIGHTENED_CLEAN_FIELD_HOST_GRP1))) {
+		vmcs12->host_ia32_pat = evmcs->host_ia32_pat;
+		vmcs12->host_ia32_efer = evmcs->host_ia32_efer;
+		vmcs12->host_cr0 = evmcs->host_cr0;
+		vmcs12->host_cr3 = evmcs->host_cr3;
+		vmcs12->host_cr4 = evmcs->host_cr4;
+		vmcs12->host_ia32_sysenter_esp = evmcs->host_ia32_sysenter_esp;
+		vmcs12->host_ia32_sysenter_eip = evmcs->host_ia32_sysenter_eip;
+		vmcs12->host_rip = evmcs->host_rip;
+		vmcs12->host_ia32_sysenter_cs = evmcs->host_ia32_sysenter_cs;
+		vmcs12->host_es_selector = evmcs->host_es_selector;
+		vmcs12->host_cs_selector = evmcs->host_cs_selector;
+		vmcs12->host_ss_selector = evmcs->host_ss_selector;
+		vmcs12->host_ds_selector = evmcs->host_ds_selector;
+		vmcs12->host_fs_selector = evmcs->host_fs_selector;
+		vmcs12->host_gs_selector = evmcs->host_gs_selector;
+		vmcs12->host_tr_selector = evmcs->host_tr_selector;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_HOST_GRP1))) {
+		vmcs12->pin_based_vm_exec_control =
+			evmcs->pin_based_vm_exec_control;
+		vmcs12->vm_exit_controls = evmcs->vm_exit_controls;
+		vmcs12->secondary_vm_exec_control =
+			evmcs->secondary_vm_exec_control;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_IO_BITMAP))) {
+		vmcs12->io_bitmap_a = evmcs->io_bitmap_a;
+		vmcs12->io_bitmap_b = evmcs->io_bitmap_b;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_MSR_BITMAP))) {
+		vmcs12->msr_bitmap = evmcs->msr_bitmap;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_GRP2))) {
+		vmcs12->guest_es_base = evmcs->guest_es_base;
+		vmcs12->guest_cs_base = evmcs->guest_cs_base;
+		vmcs12->guest_ss_base = evmcs->guest_ss_base;
+		vmcs12->guest_ds_base = evmcs->guest_ds_base;
+		vmcs12->guest_fs_base = evmcs->guest_fs_base;
+		vmcs12->guest_gs_base = evmcs->guest_gs_base;
+		vmcs12->guest_ldtr_base = evmcs->guest_ldtr_base;
+		vmcs12->guest_tr_base = evmcs->guest_tr_base;
+		vmcs12->guest_gdtr_base = evmcs->guest_gdtr_base;
+		vmcs12->guest_idtr_base = evmcs->guest_idtr_base;
+		vmcs12->guest_es_limit = evmcs->guest_es_limit;
+		vmcs12->guest_cs_limit = evmcs->guest_cs_limit;
+		vmcs12->guest_ss_limit = evmcs->guest_ss_limit;
+		vmcs12->guest_ds_limit = evmcs->guest_ds_limit;
+		vmcs12->guest_fs_limit = evmcs->guest_fs_limit;
+		vmcs12->guest_gs_limit = evmcs->guest_gs_limit;
+		vmcs12->guest_ldtr_limit = evmcs->guest_ldtr_limit;
+		vmcs12->guest_tr_limit = evmcs->guest_tr_limit;
+		vmcs12->guest_gdtr_limit = evmcs->guest_gdtr_limit;
+		vmcs12->guest_idtr_limit = evmcs->guest_idtr_limit;
+		vmcs12->guest_es_ar_bytes = evmcs->guest_es_ar_bytes;
+		vmcs12->guest_cs_ar_bytes = evmcs->guest_cs_ar_bytes;
+		vmcs12->guest_ss_ar_bytes = evmcs->guest_ss_ar_bytes;
+		vmcs12->guest_ds_ar_bytes = evmcs->guest_ds_ar_bytes;
+		vmcs12->guest_fs_ar_bytes = evmcs->guest_fs_ar_bytes;
+		vmcs12->guest_gs_ar_bytes = evmcs->guest_gs_ar_bytes;
+		vmcs12->guest_ldtr_ar_bytes = evmcs->guest_ldtr_ar_bytes;
+		vmcs12->guest_tr_ar_bytes = evmcs->guest_tr_ar_bytes;
+		vmcs12->guest_es_selector = evmcs->guest_es_selector;
+		vmcs12->guest_cs_selector = evmcs->guest_cs_selector;
+		vmcs12->guest_ss_selector = evmcs->guest_ss_selector;
+		vmcs12->guest_ds_selector = evmcs->guest_ds_selector;
+		vmcs12->guest_fs_selector = evmcs->guest_fs_selector;
+		vmcs12->guest_gs_selector = evmcs->guest_gs_selector;
+		vmcs12->guest_ldtr_selector = evmcs->guest_ldtr_selector;
+		vmcs12->guest_tr_selector = evmcs->guest_tr_selector;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_GRP2))) {
+		vmcs12->tsc_offset = evmcs->tsc_offset;
+		vmcs12->virtual_apic_page_addr = evmcs->virtual_apic_page_addr;
+		vmcs12->xss_exit_bitmap = evmcs->xss_exit_bitmap;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CRDR))) {
+		vmcs12->cr0_guest_host_mask = evmcs->cr0_guest_host_mask;
+		vmcs12->cr4_guest_host_mask = evmcs->cr4_guest_host_mask;
+		vmcs12->cr0_read_shadow = evmcs->cr0_read_shadow;
+		vmcs12->cr4_read_shadow = evmcs->cr4_read_shadow;
+		vmcs12->guest_cr0 = evmcs->guest_cr0;
+		vmcs12->guest_cr3 = evmcs->guest_cr3;
+		vmcs12->guest_cr4 = evmcs->guest_cr4;
+		vmcs12->guest_dr7 = evmcs->guest_dr7;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_HOST_POINTER))) {
+		vmcs12->host_fs_base = evmcs->host_fs_base;
+		vmcs12->host_gs_base = evmcs->host_gs_base;
+		vmcs12->host_tr_base = evmcs->host_tr_base;
+		vmcs12->host_gdtr_base = evmcs->host_gdtr_base;
+		vmcs12->host_idtr_base = evmcs->host_idtr_base;
+		vmcs12->host_rsp = evmcs->host_rsp;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_XLAT))) {
+		vmcs12->ept_pointer = evmcs->ept_pointer;
+		vmcs12->virtual_processor_id = evmcs->virtual_processor_id;
+	}
+
+	if (unlikely(full || !(evmcs->hv_clean_fields &
+			       HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_GRP1))) {
+		vmcs12->vmcs_link_pointer = evmcs->vmcs_link_pointer;
+		vmcs12->guest_ia32_debugctl = evmcs->guest_ia32_debugctl;
+		vmcs12->guest_ia32_pat = evmcs->guest_ia32_pat;
+		vmcs12->guest_ia32_efer = evmcs->guest_ia32_efer;
+		vmcs12->guest_pdptr0 = evmcs->guest_pdptr0;
+		vmcs12->guest_pdptr1 = evmcs->guest_pdptr1;
+		vmcs12->guest_pdptr2 = evmcs->guest_pdptr2;
+		vmcs12->guest_pdptr3 = evmcs->guest_pdptr3;
+		vmcs12->guest_pending_dbg_exceptions =
+			evmcs->guest_pending_dbg_exceptions;
+		vmcs12->guest_sysenter_esp = evmcs->guest_sysenter_esp;
+		vmcs12->guest_sysenter_eip = evmcs->guest_sysenter_eip;
+		vmcs12->guest_bndcfgs = evmcs->guest_bndcfgs;
+		vmcs12->guest_activity_state = evmcs->guest_activity_state;
+		vmcs12->guest_sysenter_cs = evmcs->guest_sysenter_cs;
+	}
+
+	/*
+	 * Not used?
+	 * vmcs12->vm_exit_msr_store_addr = evmcs->vm_exit_msr_store_addr;
+	 * vmcs12->vm_exit_msr_load_addr = evmcs->vm_exit_msr_load_addr;
+	 * vmcs12->vm_entry_msr_load_addr = evmcs->vm_entry_msr_load_addr;
+	 * vmcs12->cr3_target_value0 = evmcs->cr3_target_value0;
+	 * vmcs12->cr3_target_value1 = evmcs->cr3_target_value1;
+	 * vmcs12->cr3_target_value2 = evmcs->cr3_target_value2;
+	 * vmcs12->cr3_target_value3 = evmcs->cr3_target_value3;
+	 * vmcs12->page_fault_error_code_mask =
+	 *		evmcs->page_fault_error_code_mask;
+	 * vmcs12->page_fault_error_code_match =
+	 *		evmcs->page_fault_error_code_match;
+	 * vmcs12->cr3_target_count = evmcs->cr3_target_count;
+	 * vmcs12->vm_exit_msr_store_count = evmcs->vm_exit_msr_store_count;
+	 * vmcs12->vm_exit_msr_load_count = evmcs->vm_exit_msr_load_count;
+	 * vmcs12->vm_entry_msr_load_count = evmcs->vm_entry_msr_load_count;
+	 */
+
+	/*
+	 * Read only fields:
+	 * vmcs12->guest_physical_address = evmcs->guest_physical_address;
+	 * vmcs12->vm_instruction_error = evmcs->vm_instruction_error;
+	 * vmcs12->vm_exit_reason = evmcs->vm_exit_reason;
+	 * vmcs12->vm_exit_intr_info = evmcs->vm_exit_intr_info;
+	 * vmcs12->vm_exit_intr_error_code = evmcs->vm_exit_intr_error_code;
+	 * vmcs12->idt_vectoring_info_field = evmcs->idt_vectoring_info_field;
+	 * vmcs12->idt_vectoring_error_code = evmcs->idt_vectoring_error_code;
+	 * vmcs12->vm_exit_instruction_len = evmcs->vm_exit_instruction_len;
+	 * vmcs12->vmx_instruction_info = evmcs->vmx_instruction_info;
+	 * vmcs12->exit_qualification = evmcs->exit_qualification;
+	 * vmcs12->guest_linear_address = evmcs->guest_linear_address;
+	 *
+	 * Not present in struct vmcs12:
+	 * vmcs12->exit_io_instruction_ecx = evmcs->exit_io_instruction_ecx;
+	 * vmcs12->exit_io_instruction_esi = evmcs->exit_io_instruction_esi;
+	 * vmcs12->exit_io_instruction_edi = evmcs->exit_io_instruction_edi;
+	 * vmcs12->exit_io_instruction_eip = evmcs->exit_io_instruction_eip;
+	 */
+
+	return 0;
+}
+
+static int copy_vmcs12_to_enlightened(struct vcpu_vmx *vmx)
+{
+	struct vmcs12 *vmcs12 = vmx->nested.cached_vmcs12;
+	struct hv_enlightened_vmcs *evmcs = vmx->nested.hv_evmcs;
+
+	/*
+	 * Should not be changed by KVM:
+	 *
+	 * evmcs->host_es_selector = vmcs12->host_es_selector;
+	 * evmcs->host_cs_selector = vmcs12->host_cs_selector;
+	 * evmcs->host_ss_selector = vmcs12->host_ss_selector;
+	 * evmcs->host_ds_selector = vmcs12->host_ds_selector;
+	 * evmcs->host_fs_selector = vmcs12->host_fs_selector;
+	 * evmcs->host_gs_selector = vmcs12->host_gs_selector;
+	 * evmcs->host_tr_selector = vmcs12->host_tr_selector;
+	 * evmcs->host_ia32_pat = vmcs12->host_ia32_pat;
+	 * evmcs->host_ia32_efer = vmcs12->host_ia32_efer;
+	 * evmcs->host_cr0 = vmcs12->host_cr0;
+	 * evmcs->host_cr3 = vmcs12->host_cr3;
+	 * evmcs->host_cr4 = vmcs12->host_cr4;
+	 * evmcs->host_ia32_sysenter_esp = vmcs12->host_ia32_sysenter_esp;
+	 * evmcs->host_ia32_sysenter_eip = vmcs12->host_ia32_sysenter_eip;
+	 * evmcs->host_rip = vmcs12->host_rip;
+	 * evmcs->host_ia32_sysenter_cs = vmcs12->host_ia32_sysenter_cs;
+	 * evmcs->host_fs_base = vmcs12->host_fs_base;
+	 * evmcs->host_gs_base = vmcs12->host_gs_base;
+	 * evmcs->host_tr_base = vmcs12->host_tr_base;
+	 * evmcs->host_gdtr_base = vmcs12->host_gdtr_base;
+	 * evmcs->host_idtr_base = vmcs12->host_idtr_base;
+	 * evmcs->host_rsp = vmcs12->host_rsp;
+	 * sync_vmcs12() doesn't read these:
+	 * evmcs->io_bitmap_a = vmcs12->io_bitmap_a;
+	 * evmcs->io_bitmap_b = vmcs12->io_bitmap_b;
+	 * evmcs->msr_bitmap = vmcs12->msr_bitmap;
+	 * evmcs->ept_pointer = vmcs12->ept_pointer;
+	 * evmcs->xss_exit_bitmap = vmcs12->xss_exit_bitmap;
+	 * evmcs->vm_exit_msr_store_addr = vmcs12->vm_exit_msr_store_addr;
+	 * evmcs->vm_exit_msr_load_addr = vmcs12->vm_exit_msr_load_addr;
+	 * evmcs->vm_entry_msr_load_addr = vmcs12->vm_entry_msr_load_addr;
+	 * evmcs->cr3_target_value0 = vmcs12->cr3_target_value0;
+	 * evmcs->cr3_target_value1 = vmcs12->cr3_target_value1;
+	 * evmcs->cr3_target_value2 = vmcs12->cr3_target_value2;
+	 * evmcs->cr3_target_value3 = vmcs12->cr3_target_value3;
+	 * evmcs->tpr_threshold = vmcs12->tpr_threshold;
+	 * evmcs->virtual_processor_id = vmcs12->virtual_processor_id;
+	 * evmcs->exception_bitmap = vmcs12->exception_bitmap;
+	 * evmcs->vmcs_link_pointer = vmcs12->vmcs_link_pointer;
+	 * evmcs->pin_based_vm_exec_control = vmcs12->pin_based_vm_exec_control;
+	 * evmcs->vm_exit_controls = vmcs12->vm_exit_controls;
+	 * evmcs->secondary_vm_exec_control = vmcs12->secondary_vm_exec_control;
+	 * evmcs->page_fault_error_code_mask =
+	 *		vmcs12->page_fault_error_code_mask;
+	 * evmcs->page_fault_error_code_match =
+	 *		vmcs12->page_fault_error_code_match;
+	 * evmcs->cr3_target_count = vmcs12->cr3_target_count;
+	 * evmcs->virtual_apic_page_addr = vmcs12->virtual_apic_page_addr;
+	 * evmcs->tsc_offset = vmcs12->tsc_offset;
+	 * evmcs->guest_ia32_debugctl = vmcs12->guest_ia32_debugctl;
+	 * evmcs->cr0_guest_host_mask = vmcs12->cr0_guest_host_mask;
+	 * evmcs->cr4_guest_host_mask = vmcs12->cr4_guest_host_mask;
+	 * evmcs->cr0_read_shadow = vmcs12->cr0_read_shadow;
+	 * evmcs->cr4_read_shadow = vmcs12->cr4_read_shadow;
+	 * evmcs->vm_exit_msr_store_count = vmcs12->vm_exit_msr_store_count;
+	 * evmcs->vm_exit_msr_load_count = vmcs12->vm_exit_msr_load_count;
+	 * evmcs->vm_entry_msr_load_count = vmcs12->vm_entry_msr_load_count;
+	 *
+	 * Not present in struct vmcs12:
+	 * evmcs->exit_io_instruction_ecx = vmcs12->exit_io_instruction_ecx;
+	 * evmcs->exit_io_instruction_esi = vmcs12->exit_io_instruction_esi;
+	 * evmcs->exit_io_instruction_edi = vmcs12->exit_io_instruction_edi;
+	 * evmcs->exit_io_instruction_eip = vmcs12->exit_io_instruction_eip;
+	 */
+
+	evmcs->guest_es_selector = vmcs12->guest_es_selector;
+	evmcs->guest_cs_selector = vmcs12->guest_cs_selector;
+	evmcs->guest_ss_selector = vmcs12->guest_ss_selector;
+	evmcs->guest_ds_selector = vmcs12->guest_ds_selector;
+	evmcs->guest_fs_selector = vmcs12->guest_fs_selector;
+	evmcs->guest_gs_selector = vmcs12->guest_gs_selector;
+	evmcs->guest_ldtr_selector = vmcs12->guest_ldtr_selector;
+	evmcs->guest_tr_selector = vmcs12->guest_tr_selector;
+
+	evmcs->guest_es_limit = vmcs12->guest_es_limit;
+	evmcs->guest_cs_limit = vmcs12->guest_cs_limit;
+	evmcs->guest_ss_limit = vmcs12->guest_ss_limit;
+	evmcs->guest_ds_limit = vmcs12->guest_ds_limit;
+	evmcs->guest_fs_limit = vmcs12->guest_fs_limit;
+	evmcs->guest_gs_limit = vmcs12->guest_gs_limit;
+	evmcs->guest_ldtr_limit = vmcs12->guest_ldtr_limit;
+	evmcs->guest_tr_limit = vmcs12->guest_tr_limit;
+	evmcs->guest_gdtr_limit = vmcs12->guest_gdtr_limit;
+	evmcs->guest_idtr_limit = vmcs12->guest_idtr_limit;
+
+	evmcs->guest_es_ar_bytes = vmcs12->guest_es_ar_bytes;
+	evmcs->guest_cs_ar_bytes = vmcs12->guest_cs_ar_bytes;
+	evmcs->guest_ss_ar_bytes = vmcs12->guest_ss_ar_bytes;
+	evmcs->guest_ds_ar_bytes = vmcs12->guest_ds_ar_bytes;
+	evmcs->guest_fs_ar_bytes = vmcs12->guest_fs_ar_bytes;
+	evmcs->guest_gs_ar_bytes = vmcs12->guest_gs_ar_bytes;
+	evmcs->guest_ldtr_ar_bytes = vmcs12->guest_ldtr_ar_bytes;
+	evmcs->guest_tr_ar_bytes = vmcs12->guest_tr_ar_bytes;
+
+	evmcs->guest_es_base = vmcs12->guest_es_base;
+	evmcs->guest_cs_base = vmcs12->guest_cs_base;
+	evmcs->guest_ss_base = vmcs12->guest_ss_base;
+	evmcs->guest_ds_base = vmcs12->guest_ds_base;
+	evmcs->guest_fs_base = vmcs12->guest_fs_base;
+	evmcs->guest_gs_base = vmcs12->guest_gs_base;
+	evmcs->guest_ldtr_base = vmcs12->guest_ldtr_base;
+	evmcs->guest_tr_base = vmcs12->guest_tr_base;
+	evmcs->guest_gdtr_base = vmcs12->guest_gdtr_base;
+	evmcs->guest_idtr_base = vmcs12->guest_idtr_base;
+
+	evmcs->guest_ia32_pat = vmcs12->guest_ia32_pat;
+	evmcs->guest_ia32_efer = vmcs12->guest_ia32_efer;
+
+	evmcs->guest_pdptr0 = vmcs12->guest_pdptr0;
+	evmcs->guest_pdptr1 = vmcs12->guest_pdptr1;
+	evmcs->guest_pdptr2 = vmcs12->guest_pdptr2;
+	evmcs->guest_pdptr3 = vmcs12->guest_pdptr3;
+
+	evmcs->guest_pending_dbg_exceptions =
+		vmcs12->guest_pending_dbg_exceptions;
+	evmcs->guest_sysenter_esp = vmcs12->guest_sysenter_esp;
+	evmcs->guest_sysenter_eip = vmcs12->guest_sysenter_eip;
+
+	evmcs->guest_activity_state = vmcs12->guest_activity_state;
+	evmcs->guest_sysenter_cs = vmcs12->guest_sysenter_cs;
+
+	evmcs->guest_cr0 = vmcs12->guest_cr0;
+	evmcs->guest_cr3 = vmcs12->guest_cr3;
+	evmcs->guest_cr4 = vmcs12->guest_cr4;
+	evmcs->guest_dr7 = vmcs12->guest_dr7;
+
+	evmcs->guest_physical_address = vmcs12->guest_physical_address;
+
+	evmcs->vm_instruction_error = vmcs12->vm_instruction_error;
+	evmcs->vm_exit_reason = vmcs12->vm_exit_reason;
+	evmcs->vm_exit_intr_info = vmcs12->vm_exit_intr_info;
+	evmcs->vm_exit_intr_error_code = vmcs12->vm_exit_intr_error_code;
+	evmcs->idt_vectoring_info_field = vmcs12->idt_vectoring_info_field;
+	evmcs->idt_vectoring_error_code = vmcs12->idt_vectoring_error_code;
+	evmcs->vm_exit_instruction_len = vmcs12->vm_exit_instruction_len;
+	evmcs->vmx_instruction_info = vmcs12->vmx_instruction_info;
+
+	evmcs->exit_qualification = vmcs12->exit_qualification;
+
+	evmcs->guest_linear_address = vmcs12->guest_linear_address;
+	evmcs->guest_rsp = vmcs12->guest_rsp;
+	evmcs->guest_rflags = vmcs12->guest_rflags;
+
+	evmcs->guest_interruptibility_info =
+		vmcs12->guest_interruptibility_info;
+	evmcs->cpu_based_vm_exec_control = vmcs12->cpu_based_vm_exec_control;
+	evmcs->vm_entry_controls = vmcs12->vm_entry_controls;
+	evmcs->vm_entry_intr_info_field = vmcs12->vm_entry_intr_info_field;
+	evmcs->vm_entry_exception_error_code =
+		vmcs12->vm_entry_exception_error_code;
+	evmcs->vm_entry_instruction_len = vmcs12->vm_entry_instruction_len;
+
+	evmcs->guest_rip = vmcs12->guest_rip;
+
+	evmcs->guest_bndcfgs = vmcs12->guest_bndcfgs;
+
+	return 0;
+}
+
 /*
  * Copy the writable VMCS shadow fields back to the VMCS12, in case
  * they have been modified by the L1 guest. Note that the "read-only"
@@ -8403,7 +8792,7 @@ static void set_current_vmptr(struct vcpu_vmx *vmx, gpa_t vmptr)
 			      SECONDARY_EXEC_SHADOW_VMCS);
 		vmcs_write64(VMCS_LINK_POINTER,
 			     __pa(vmx->vmcs01.shadow_vmcs));
-		vmx->nested.sync_shadow_vmcs = true;
+		vmx->nested.need_vmcs12_sync = true;
 	}
 	vmx->nested.dirty_vmcs12 = true;
 }
@@ -9977,9 +10366,16 @@ static void __noclone vmx_vcpu_run(struct kvm_vcpu *vcpu)
 		vmcs_write32(PLE_WINDOW, vmx->ple_window);
 	}
 
-	if (vmx->nested.sync_shadow_vmcs) {
-		copy_vmcs12_to_shadow(vmx);
-		vmx->nested.sync_shadow_vmcs = false;
+	if (vmx->nested.need_vmcs12_sync) {
+		if (vmx->nested.hv_evmcs) {
+			copy_vmcs12_to_enlightened(vmx);
+			/* All fields are clean */
+			vmx->nested.hv_evmcs->hv_clean_fields |=
+				HV_VMX_ENLIGHTENED_CLEAN_FIELD_ALL;
+		} else {
+			copy_vmcs12_to_shadow(vmx);
+		}
+		vmx->nested.need_vmcs12_sync = false;
 	}
 
 	if (test_bit(VCPU_REGS_RSP, (unsigned long *)&vcpu->arch.regs_dirty))
@@ -11296,7 +11692,7 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12,
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	u32 exec_control, vmcs12_exec_ctrl;
 
-	if (vmx->nested.dirty_vmcs12) {
+	if (vmx->nested.dirty_vmcs12 || vmx->nested.hv_evmcs) {
 		prepare_vmcs02_full(vcpu, vmcs12);
 		vmx->nested.dirty_vmcs12 = false;
 	}
@@ -11772,8 +12168,13 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
 
 	vmcs12 = get_vmcs12(vcpu);
 
-	if (enable_shadow_vmcs)
+	if (vmx->nested.hv_evmcs) {
+		copy_enlightened_to_vmcs12(vmx, vmx->nested.dirty_vmcs12);
+		/* Enlightened VMCS doesn't have launch state */
+		vmcs12->launch_state = !launch;
+	} else if (enable_shadow_vmcs) {
 		copy_shadow_to_vmcs12(vmx);
+	}
 
 	/*
 	 * The nested entry process starts with enforcing various prerequisites
@@ -12398,8 +12799,8 @@ static void nested_vmx_vmexit(struct kvm_vcpu *vcpu, u32 exit_reason,
 	 */
 	kvm_make_request(KVM_REQ_APIC_PAGE_RELOAD, vcpu);
 
-	if (enable_shadow_vmcs && exit_reason != -1)
-		vmx->nested.sync_shadow_vmcs = true;
+	if ((exit_reason != -1) && (enable_shadow_vmcs || vmx->nested.hv_evmcs))
+		vmx->nested.need_vmcs12_sync = true;
 
 	/* in case we halted in L2 */
 	vcpu->arch.mp_state = KVM_MP_STATE_RUNNABLE;
@@ -12478,12 +12879,14 @@ static void nested_vmx_entry_failure(struct kvm_vcpu *vcpu,
 			struct vmcs12 *vmcs12,
 			u32 reason, unsigned long qualification)
 {
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
+
 	load_vmcs12_host_state(vcpu, vmcs12);
 	vmcs12->vm_exit_reason = reason | VMX_EXIT_REASONS_FAILED_VMENTRY;
 	vmcs12->exit_qualification = qualification;
 	nested_vmx_succeed(vcpu);
-	if (enable_shadow_vmcs)
-		to_vmx(vcpu)->nested.sync_shadow_vmcs = true;
+	if (enable_shadow_vmcs || vmx->nested.hv_evmcs)
+		vmx->nested.need_vmcs12_sync = true;
 }
 
 static int vmx_check_intercept(struct kvm_vcpu *vcpu,
-- 
2.14.4

[PATCH v2 5/6] KVM: nVMX: implement enlightened VMPTRLD and VMCLEAR

[Vitaly Kuznetsov]

Per Hyper-V TLFS 5.0b:

"The L1 hypervisor may choose to use enlightened VMCSs by writing 1 to
the corresponding field in the VP assist page (see section 7.8.7).
Another field in the VP assist page controls the currently active
enlightened VMCS. Each enlightened VMCS is exactly one page (4 KB) in
size and must be initially zeroed. No VMPTRLD instruction must be
executed to make an enlightened VMCS active or current.

After the L1 hypervisor performs a VM entry with an enlightened VMCS,
the VMCS is considered active on the processor. An enlightened VMCS
can only be active on a single processor at the same time. The L1
hypervisor can execute a VMCLEAR instruction to transition an
enlightened VMCS from the active to the non-active state. Any VMREAD
or VMWRITE instructions while an enlightened VMCS is active is
unsupported and can result in unexpected behavior."

Keep Enlightened VMCS structure for the current L2 guest permanently mapped
from struct nested_vmx instead of mapping it every time.

Suggested-by: Ladi Prosek <lprosek@redhat.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Liran Alon <liran.alon@oracle.com>
---
 arch/x86/kvm/vmx.c | 98 ++++++++++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 91 insertions(+), 7 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 47c34c2c3ba7..ea6c651ed477 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -20,6 +20,7 @@
 #include "mmu.h"
 #include "cpuid.h"
 #include "lapic.h"
+#include "hyperv.h"
 
 #include <linux/kvm_host.h>
 #include <linux/module.h>
@@ -690,6 +691,8 @@ struct nested_vmx {
 		bool guest_mode;
 	} smm;
 
+	gpa_t hv_evmcs_vmptr;
+	struct page *hv_evmcs_page;
 	struct hv_enlightened_vmcs *hv_evmcs;
 };
 
@@ -7700,7 +7703,9 @@ static void nested_vmx_failInvalid(struct kvm_vcpu *vcpu)
 static void nested_vmx_failValid(struct kvm_vcpu *vcpu,
 					u32 vm_instruction_error)
 {
-	if (to_vmx(vcpu)->nested.current_vmptr == -1ull) {
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
+
+	if (vmx->nested.current_vmptr == -1ull && !vmx->nested.hv_evmcs) {
 		/*
 		 * failValid writes the error number to the current VMCS, which
 		 * can't be done there isn't a current VMCS.
@@ -8008,6 +8013,18 @@ static void vmx_disable_shadow_vmcs(struct vcpu_vmx *vmx)
 	vmcs_write64(VMCS_LINK_POINTER, -1ull);
 }
 
+static inline void nested_release_evmcs(struct vcpu_vmx *vmx)
+{
+	if (!vmx->nested.hv_evmcs)
+		return;
+
+	kunmap(vmx->nested.hv_evmcs_page);
+	kvm_release_page_dirty(vmx->nested.hv_evmcs_page);
+	vmx->nested.hv_evmcs_vmptr = -1ull;
+	vmx->nested.hv_evmcs_page = NULL;
+	vmx->nested.hv_evmcs = NULL;
+}
+
 static inline void nested_release_vmcs12(struct vcpu_vmx *vmx)
 {
 	if (vmx->nested.current_vmptr == -1ull)
@@ -8067,6 +8084,8 @@ static void free_nested(struct vcpu_vmx *vmx)
 		vmx->nested.pi_desc = NULL;
 	}
 
+	nested_release_evmcs(vmx);
+
 	free_loaded_vmcs(&vmx->nested.vmcs02);
 }
 
@@ -8103,12 +8122,18 @@ static int handle_vmclear(struct kvm_vcpu *vcpu)
 		return kvm_skip_emulated_instruction(vcpu);
 	}
 
-	if (vmptr == vmx->nested.current_vmptr)
-		nested_release_vmcs12(vmx);
+	if (vmx->nested.hv_evmcs_page) {
+		if (vmptr == vmx->nested.hv_evmcs_vmptr)
+			nested_release_evmcs(vmx);
+	} else {
+		if (vmptr == vmx->nested.current_vmptr)
+			nested_release_vmcs12(vmx);
 
-	kvm_vcpu_write_guest(vcpu,
-			vmptr + offsetof(struct vmcs12, launch_state),
-			&zero, sizeof(zero));
+		kvm_vcpu_write_guest(vcpu,
+				     vmptr + offsetof(struct vmcs12,
+						      launch_state),
+				     &zero, sizeof(zero));
+	}
 
 	nested_vmx_succeed(vcpu);
 	return kvm_skip_emulated_instruction(vcpu);
@@ -8819,6 +8844,10 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
 		return kvm_skip_emulated_instruction(vcpu);
 	}
 
+	/* Forbid normal VMPTRLD if Enlightened version was used */
+	if (vmx->nested.hv_evmcs)
+		return 1;
+
 	if (vmx->nested.current_vmptr != vmptr) {
 		struct vmcs12 *new_vmcs12;
 		struct page *page;
@@ -8852,6 +8881,55 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
 	return kvm_skip_emulated_instruction(vcpu);
 }
 
+/*
+ * This is an equivalent of the nested hypervisor executing the vmptrld
+ * instruction.
+ */
+static int nested_vmx_handle_enlightened_vmptrld(struct kvm_vcpu *vcpu)
+{
+	struct vcpu_vmx *vmx = to_vmx(vcpu);
+	struct hv_vp_assist_page assist_page;
+
+	if (likely(!vmx->nested.enlightened_vmcs_enabled))
+		return 1;
+
+	if (unlikely(!kvm_hv_get_assist_page(vcpu, &assist_page)))
+		return 1;
+
+	if (unlikely(!assist_page.enlighten_vmentry))
+		return 1;
+
+	if (unlikely(assist_page.current_nested_vmcs !=
+		     vmx->nested.hv_evmcs_vmptr)) {
+
+		if (!vmx->nested.hv_evmcs)
+			vmx->nested.current_vmptr = -1ull;
+
+		nested_release_evmcs(vmx);
+
+		vmx->nested.hv_evmcs_page = kvm_vcpu_gpa_to_page(
+			vcpu, assist_page.current_nested_vmcs);
+
+		if (unlikely(is_error_page(vmx->nested.hv_evmcs_page)))
+			return 0;
+
+		vmx->nested.hv_evmcs = kmap(vmx->nested.hv_evmcs_page);
+		vmx->nested.dirty_vmcs12 = true;
+		vmx->nested.hv_evmcs_vmptr = assist_page.current_nested_vmcs;
+
+		/*
+		 * Unlike normal vmcs12, enlightened vmcs12 is not fully
+		 * reloaded from guest's memory (read only fields, fields not
+		 * present in struct hv_enlightened_vmcs, ...). Make sure there
+		 * are no leftovers.
+		 */
+		memset(vmx->nested.cached_vmcs12, 0,
+		       sizeof(*vmx->nested.cached_vmcs12));
+
+	}
+	return 1;
+}
+
 /* Emulate the VMPTRST instruction */
 static int handle_vmptrst(struct kvm_vcpu *vcpu)
 {
@@ -8863,6 +8941,9 @@ static int handle_vmptrst(struct kvm_vcpu *vcpu)
 	if (!nested_vmx_check_permission(vcpu))
 		return 1;
 
+	if (unlikely(to_vmx(vcpu)->nested.hv_evmcs))
+		return 1;
+
 	if (get_vmx_mem_address(vcpu, exit_qualification,
 			vmx_instruction_info, true, &vmcs_gva))
 		return 1;
@@ -12163,7 +12244,10 @@ static int nested_vmx_run(struct kvm_vcpu *vcpu, bool launch)
 	if (!nested_vmx_check_permission(vcpu))
 		return 1;
 
-	if (!nested_vmx_check_vmcs12(vcpu))
+	if (!nested_vmx_handle_enlightened_vmptrld(vcpu))
+		return 1;
+
+	if (!vmx->nested.hv_evmcs && !nested_vmx_check_vmcs12(vcpu))
 		goto out;
 
 	vmcs12 = get_vmcs12(vcpu);
-- 
2.14.4

[PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

When Enlightened VMCS is in use by L1 hypervisor we can avoid vmwriting
VMCS fields which did not change.

Our first goal is to achieve minimal impact on traditional VMCS case so
we're not wrapping each vmwrite() with an if-changed checker. We also can't
utilize static keys as Enlightened VMCS usage is per-guest.

This patch implements the simpliest solution: checking fields in groups.
We skip single vmwrite() statements as doing the check will cost us
something even in non-evmcs case and the win is tiny. Unfortunately, this
makes prepare_vmcs02_full{,_full}() code Enlightened VMCS-dependent (and
a bit ugly).

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
Changes since v1:
- Convert evmcs_needs_write() to inline function [Liran Alon]
---
 arch/x86/kvm/vmx.c | 158 +++++++++++++++++++++++++++++++----------------------
 1 file changed, 94 insertions(+), 64 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index ea6c651ed477..857fa86c6918 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -11634,47 +11634,82 @@ static int nested_vmx_load_cr3(struct kvm_vcpu *vcpu, unsigned long cr3, bool ne
 	return 0;
 }
 
+/*
+ * Check if L1 hypervisor changed the particular field in Enlightened
+ * VMCS and avoid redundant vmwrite if it didn't. Can only be used when
+ * the value we're about to write is unchanged vmcs12->field.
+ */
+static __always_inline bool evmcs_needs_write(struct vcpu_vmx *vmx,
+					      u16 clean_field)
+{
+	if (vmx->nested.dirty_vmcs12)
+		return true;
+
+	return !(vmx->nested.hv_evmcs->hv_clean_fields & clean_field);
+}
+
 static void prepare_vmcs02_full(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
+	struct hv_enlightened_vmcs *hv_evmcs = vmx->nested.hv_evmcs;
+
+	if (!hv_evmcs ||
+	    evmcs_needs_write(vmx, HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_GRP2)) {
+		vmcs_write16(GUEST_ES_SELECTOR, vmcs12->guest_es_selector);
+		vmcs_write16(GUEST_SS_SELECTOR, vmcs12->guest_ss_selector);
+		vmcs_write16(GUEST_DS_SELECTOR, vmcs12->guest_ds_selector);
+		vmcs_write16(GUEST_FS_SELECTOR, vmcs12->guest_fs_selector);
+		vmcs_write16(GUEST_GS_SELECTOR, vmcs12->guest_gs_selector);
+		vmcs_write16(GUEST_LDTR_SELECTOR, vmcs12->guest_ldtr_selector);
+		vmcs_write16(GUEST_TR_SELECTOR, vmcs12->guest_tr_selector);
+		vmcs_write32(GUEST_ES_LIMIT, vmcs12->guest_es_limit);
+		vmcs_write32(GUEST_SS_LIMIT, vmcs12->guest_ss_limit);
+		vmcs_write32(GUEST_DS_LIMIT, vmcs12->guest_ds_limit);
+		vmcs_write32(GUEST_FS_LIMIT, vmcs12->guest_fs_limit);
+		vmcs_write32(GUEST_GS_LIMIT, vmcs12->guest_gs_limit);
+		vmcs_write32(GUEST_LDTR_LIMIT, vmcs12->guest_ldtr_limit);
+		vmcs_write32(GUEST_TR_LIMIT, vmcs12->guest_tr_limit);
+		vmcs_write32(GUEST_GDTR_LIMIT, vmcs12->guest_gdtr_limit);
+		vmcs_write32(GUEST_IDTR_LIMIT, vmcs12->guest_idtr_limit);
+		vmcs_write32(GUEST_ES_AR_BYTES, vmcs12->guest_es_ar_bytes);
+		vmcs_write32(GUEST_SS_AR_BYTES, vmcs12->guest_ss_ar_bytes);
+		vmcs_write32(GUEST_DS_AR_BYTES, vmcs12->guest_ds_ar_bytes);
+		vmcs_write32(GUEST_FS_AR_BYTES, vmcs12->guest_fs_ar_bytes);
+		vmcs_write32(GUEST_GS_AR_BYTES, vmcs12->guest_gs_ar_bytes);
+		vmcs_write32(GUEST_LDTR_AR_BYTES, vmcs12->guest_ldtr_ar_bytes);
+		vmcs_write32(GUEST_TR_AR_BYTES, vmcs12->guest_tr_ar_bytes);
+		vmcs_writel(GUEST_SS_BASE, vmcs12->guest_ss_base);
+		vmcs_writel(GUEST_DS_BASE, vmcs12->guest_ds_base);
+		vmcs_writel(GUEST_FS_BASE, vmcs12->guest_fs_base);
+		vmcs_writel(GUEST_GS_BASE, vmcs12->guest_gs_base);
+		vmcs_writel(GUEST_LDTR_BASE, vmcs12->guest_ldtr_base);
+		vmcs_writel(GUEST_TR_BASE, vmcs12->guest_tr_base);
+		vmcs_writel(GUEST_GDTR_BASE, vmcs12->guest_gdtr_base);
+		vmcs_writel(GUEST_IDTR_BASE, vmcs12->guest_idtr_base);
+	}
+
+	if (!hv_evmcs ||
+	    evmcs_needs_write(vmx, HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_GRP1)) {
+		vmcs_write32(GUEST_SYSENTER_CS, vmcs12->guest_sysenter_cs);
+		vmcs_writel(GUEST_PENDING_DBG_EXCEPTIONS,
+			    vmcs12->guest_pending_dbg_exceptions);
+		vmcs_writel(GUEST_SYSENTER_ESP, vmcs12->guest_sysenter_esp);
+		vmcs_writel(GUEST_SYSENTER_EIP, vmcs12->guest_sysenter_eip);
+
+		if (vmx_mpx_supported())
+			vmcs_write64(GUEST_BNDCFGS, vmcs12->guest_bndcfgs);
 
-	vmcs_write16(GUEST_ES_SELECTOR, vmcs12->guest_es_selector);
-	vmcs_write16(GUEST_SS_SELECTOR, vmcs12->guest_ss_selector);
-	vmcs_write16(GUEST_DS_SELECTOR, vmcs12->guest_ds_selector);
-	vmcs_write16(GUEST_FS_SELECTOR, vmcs12->guest_fs_selector);
-	vmcs_write16(GUEST_GS_SELECTOR, vmcs12->guest_gs_selector);
-	vmcs_write16(GUEST_LDTR_SELECTOR, vmcs12->guest_ldtr_selector);
-	vmcs_write16(GUEST_TR_SELECTOR, vmcs12->guest_tr_selector);
-	vmcs_write32(GUEST_ES_LIMIT, vmcs12->guest_es_limit);
-	vmcs_write32(GUEST_SS_LIMIT, vmcs12->guest_ss_limit);
-	vmcs_write32(GUEST_DS_LIMIT, vmcs12->guest_ds_limit);
-	vmcs_write32(GUEST_FS_LIMIT, vmcs12->guest_fs_limit);
-	vmcs_write32(GUEST_GS_LIMIT, vmcs12->guest_gs_limit);
-	vmcs_write32(GUEST_LDTR_LIMIT, vmcs12->guest_ldtr_limit);
-	vmcs_write32(GUEST_TR_LIMIT, vmcs12->guest_tr_limit);
-	vmcs_write32(GUEST_GDTR_LIMIT, vmcs12->guest_gdtr_limit);
-	vmcs_write32(GUEST_IDTR_LIMIT, vmcs12->guest_idtr_limit);
-	vmcs_write32(GUEST_ES_AR_BYTES, vmcs12->guest_es_ar_bytes);
-	vmcs_write32(GUEST_SS_AR_BYTES, vmcs12->guest_ss_ar_bytes);
-	vmcs_write32(GUEST_DS_AR_BYTES, vmcs12->guest_ds_ar_bytes);
-	vmcs_write32(GUEST_FS_AR_BYTES, vmcs12->guest_fs_ar_bytes);
-	vmcs_write32(GUEST_GS_AR_BYTES, vmcs12->guest_gs_ar_bytes);
-	vmcs_write32(GUEST_LDTR_AR_BYTES, vmcs12->guest_ldtr_ar_bytes);
-	vmcs_write32(GUEST_TR_AR_BYTES, vmcs12->guest_tr_ar_bytes);
-	vmcs_writel(GUEST_SS_BASE, vmcs12->guest_ss_base);
-	vmcs_writel(GUEST_DS_BASE, vmcs12->guest_ds_base);
-	vmcs_writel(GUEST_FS_BASE, vmcs12->guest_fs_base);
-	vmcs_writel(GUEST_GS_BASE, vmcs12->guest_gs_base);
-	vmcs_writel(GUEST_LDTR_BASE, vmcs12->guest_ldtr_base);
-	vmcs_writel(GUEST_TR_BASE, vmcs12->guest_tr_base);
-	vmcs_writel(GUEST_GDTR_BASE, vmcs12->guest_gdtr_base);
-	vmcs_writel(GUEST_IDTR_BASE, vmcs12->guest_idtr_base);
-
-	vmcs_write32(GUEST_SYSENTER_CS, vmcs12->guest_sysenter_cs);
-	vmcs_writel(GUEST_PENDING_DBG_EXCEPTIONS,
-		vmcs12->guest_pending_dbg_exceptions);
-	vmcs_writel(GUEST_SYSENTER_ESP, vmcs12->guest_sysenter_esp);
-	vmcs_writel(GUEST_SYSENTER_EIP, vmcs12->guest_sysenter_eip);
+		/*
+		 * L1 may access the L2's PDPTR, so save them to construct
+		 * vmcs12
+		 */
+		if (enable_ept) {
+			vmcs_write64(GUEST_PDPTR0, vmcs12->guest_pdptr0);
+			vmcs_write64(GUEST_PDPTR1, vmcs12->guest_pdptr1);
+			vmcs_write64(GUEST_PDPTR2, vmcs12->guest_pdptr2);
+			vmcs_write64(GUEST_PDPTR3, vmcs12->guest_pdptr3);
+		}
+	}
 
 	if (nested_cpu_has_xsaves(vmcs12))
 		vmcs_write64(XSS_EXIT_BITMAP, vmcs12->xss_exit_bitmap);
@@ -11732,9 +11767,6 @@ static void prepare_vmcs02_full(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 
 	set_cr4_guest_host_mask(vmx);
 
-	if (vmx_mpx_supported())
-		vmcs_write64(GUEST_BNDCFGS, vmcs12->guest_bndcfgs);
-
 	if (enable_vpid) {
 		if (nested_cpu_has_vpid(vmcs12) && vmx->nested.vpid02)
 			vmcs_write16(VIRTUAL_PROCESSOR_ID, vmx->nested.vpid02);
@@ -11742,16 +11774,6 @@ static void prepare_vmcs02_full(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 			vmcs_write16(VIRTUAL_PROCESSOR_ID, vmx->vpid);
 	}
 
-	/*
-	 * L1 may access the L2's PDPTR, so save them to construct vmcs12
-	 */
-	if (enable_ept) {
-		vmcs_write64(GUEST_PDPTR0, vmcs12->guest_pdptr0);
-		vmcs_write64(GUEST_PDPTR1, vmcs12->guest_pdptr1);
-		vmcs_write64(GUEST_PDPTR2, vmcs12->guest_pdptr2);
-		vmcs_write64(GUEST_PDPTR3, vmcs12->guest_pdptr3);
-	}
-
 	if (cpu_has_vmx_msr_bitmap())
 		vmcs_write64(MSR_BITMAP, __pa(vmx->nested.vmcs02.msr_bitmap));
 }
@@ -11772,22 +11794,24 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12,
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	u32 exec_control, vmcs12_exec_ctrl;
+	struct hv_enlightened_vmcs *hv_evmcs = vmx->nested.hv_evmcs;
 
-	if (vmx->nested.dirty_vmcs12 || vmx->nested.hv_evmcs) {
+	if (vmx->nested.dirty_vmcs12 || vmx->nested.hv_evmcs)
 		prepare_vmcs02_full(vcpu, vmcs12);
-		vmx->nested.dirty_vmcs12 = false;
-	}
 
 	/*
 	 * First, the fields that are shadowed.  This must be kept in sync
 	 * with vmx_shadow_fields.h.
 	 */
 
-	vmcs_write16(GUEST_CS_SELECTOR, vmcs12->guest_cs_selector);
-	vmcs_write32(GUEST_CS_LIMIT, vmcs12->guest_cs_limit);
-	vmcs_write32(GUEST_CS_AR_BYTES, vmcs12->guest_cs_ar_bytes);
-	vmcs_writel(GUEST_ES_BASE, vmcs12->guest_es_base);
-	vmcs_writel(GUEST_CS_BASE, vmcs12->guest_cs_base);
+	if (!hv_evmcs ||
+	    evmcs_needs_write(vmx, HV_VMX_ENLIGHTENED_CLEAN_FIELD_GUEST_GRP2)) {
+		vmcs_write16(GUEST_CS_SELECTOR, vmcs12->guest_cs_selector);
+		vmcs_write32(GUEST_CS_LIMIT, vmcs12->guest_cs_limit);
+		vmcs_write32(GUEST_CS_AR_BYTES, vmcs12->guest_cs_ar_bytes);
+		vmcs_writel(GUEST_ES_BASE, vmcs12->guest_es_base);
+		vmcs_writel(GUEST_CS_BASE, vmcs12->guest_cs_base);
+	}
 
 	/*
 	 * Not in vmcs02: GUEST_PML_INDEX, HOST_FS_SELECTOR, HOST_GS_SELECTOR,
@@ -11803,12 +11827,15 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12,
 		vmcs_write64(GUEST_IA32_DEBUGCTL, vmx->nested.vmcs01_debugctl);
 	}
 	if (vmx->nested.nested_run_pending) {
-		vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
-			     vmcs12->vm_entry_intr_info_field);
-		vmcs_write32(VM_ENTRY_EXCEPTION_ERROR_CODE,
-			     vmcs12->vm_entry_exception_error_code);
-		vmcs_write32(VM_ENTRY_INSTRUCTION_LEN,
-			     vmcs12->vm_entry_instruction_len);
+		if (!hv_evmcs || evmcs_needs_write(vmx,
+			   HV_VMX_ENLIGHTENED_CLEAN_FIELD_CONTROL_EVENT)) {
+			vmcs_write32(VM_ENTRY_INTR_INFO_FIELD,
+				     vmcs12->vm_entry_intr_info_field);
+			vmcs_write32(VM_ENTRY_EXCEPTION_ERROR_CODE,
+				     vmcs12->vm_entry_exception_error_code);
+			vmcs_write32(VM_ENTRY_INSTRUCTION_LEN,
+				     vmcs12->vm_entry_instruction_len);
+		}
 		vmcs_write32(GUEST_INTERRUPTIBILITY_INFO,
 			     vmcs12->guest_interruptibility_info);
 		vmx->loaded_vmcs->nmi_known_unmasked =
@@ -12032,6 +12059,9 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12,
 
 	kvm_register_write(vcpu, VCPU_REGS_RSP, vmcs12->guest_rsp);
 	kvm_register_write(vcpu, VCPU_REGS_RIP, vmcs12->guest_rip);
+
+	vmx->nested.dirty_vmcs12 = false;
+
 	return 0;
 }
 
-- 
2.14.4

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Paolo Bonzini]

On 21/06/2018 14:30, Vitaly Kuznetsov wrote:
> +static __always_inline bool evmcs_needs_write(struct vcpu_vmx *vmx,
> +					      u16 clean_field)
> +{
> +	if (vmx->nested.dirty_vmcs12)
> +		return true;

Why is this needed?  If it weren't for it, you could pass hv_evmcs
directly to evmcs_needs_write, which would simplify the code a bit in
the caller.

Paolo

> +	return !(vmx->nested.hv_evmcs->hv_clean_fields & clean_field);

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 21/06/2018 14:30, Vitaly Kuznetsov wrote:
>> +static __always_inline bool evmcs_needs_write(struct vcpu_vmx *vmx,
>> +					      u16 clean_field)
>> +{
>> +	if (vmx->nested.dirty_vmcs12)
>> +		return true;
>
> Why is this needed?  If it weren't for it, you could pass hv_evmcs
> directly to evmcs_needs_write, which would simplify the code a bit in
> the caller.

This is an equivalent of prepare_vmcs02()/prepare_vmcs02_full() split
for eVMCS case: when we switch from L2 guest A to L2 guest B we need to
write the whole VMCS so evmcs_needs_write() needs to return true.

We can, however, make an optimisation: forcefuly reset hv_clean_fields
mask on enlightened vmptrld making 'dirty_vmcs12' check redundant.

This series was posted quite some time ago, I'll rebase and re-test with
the fix included.

Thanks!

-- 
  Vitaly

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Paolo Bonzini]

On 25/07/2018 10:37, Vitaly Kuznetsov wrote:
>> Why is this needed?  If it weren't for it, you could pass hv_evmcs
>> directly to evmcs_needs_write, which would simplify the code a bit in
>> the caller.
> This is an equivalent of prepare_vmcs02()/prepare_vmcs02_full() split
> for eVMCS case: when we switch from L2 guest A to L2 guest B we need to
> write the whole VMCS so evmcs_needs_write() needs to return true.

Right, I missed the dirty_vmcs12 assignment in patch 5.

But is L0 allowed to write to hv_clean_fields?  One possibility is to
add a dirty_evmcs field to struct nested_vmx, and "OR" ~hv_clean_fields
into it at the beginning of prepare_vmcs02.

Something like

	if (vmx->nested.hv_evmcs) {
		vmx->nested.dirty_evmcs |=
			~vmx->nested.hv_evmcs->hv_clean_fields;
		prepare_vmcs02_full(vcpu, vmcs12,
				    vmx->nested.dirty_evmcs);
	} else if (vmx->nested.dirty_vmcs12) {
		prepare_vmcs02_full(vcpu, vmcs12, ~0);
	}

	...
	vmx->nested.dirty_evmcs = 0;
	vmx->nested.dirty_vmcs12 = false;

?

Paolo

> We can, however, make an optimisation: forcefuly reset hv_clean_fields
> mask on enlightened vmptrld making 'dirty_vmcs12' check redundant.

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 25/07/2018 10:37, Vitaly Kuznetsov wrote:
>>> Why is this needed?  If it weren't for it, you could pass hv_evmcs
>>> directly to evmcs_needs_write, which would simplify the code a bit in
>>> the caller.
>> This is an equivalent of prepare_vmcs02()/prepare_vmcs02_full() split
>> for eVMCS case: when we switch from L2 guest A to L2 guest B we need to
>> write the whole VMCS so evmcs_needs_write() needs to return true.
>
> Right, I missed the dirty_vmcs12 assignment in patch 5.
>
> But is L0 allowed to write to hv_clean_fields?  

It is kinda expected to: currently I reset it in vmx_vcpu_run() and (if
I remember correctly) L1 Hyper-V only clears bits in this mask when it
touches certain fields so if we don't set it to 'all clean' it stays
zeroed forever. So nothing stops us from doing 

       if (hv_evmcs && vmx->nested.dirty_vmcs12)
               hv_evmcs->hv_clean_fields &=
                       ~HV_VMX_ENLIGHTENED_CLEAN_FIELD_ALL;

in prepare_vmcs02() I guess.

> One possibility is to
> add a dirty_evmcs field to struct nested_vmx, and "OR" ~hv_clean_fields
> into it at the beginning of prepare_vmcs02.
>
> Something like
>
> 	if (vmx->nested.hv_evmcs) {
> 		vmx->nested.dirty_evmcs |=
> 			~vmx->nested.hv_evmcs->hv_clean_fields;
> 		prepare_vmcs02_full(vcpu, vmcs12,
> 				    vmx->nested.dirty_evmcs);
> 	} else if (vmx->nested.dirty_vmcs12) {
> 		prepare_vmcs02_full(vcpu, vmcs12, ~0);
> 	}
>
> 	...
> 	vmx->nested.dirty_evmcs = 0;
> 	vmx->nested.dirty_vmcs12 = false;
>
> ?
>

I think we can even get away with a local variable in prepare_vmcs02()
and pass it to prepare_vmcs02_full(), no need to have it in struct
nested_vmx. But I would slightly prefer to just reset
hv_evmcs->hv_clean_fields when vmcs12 is dirty.

Thanks,

-- 
  Vitaly

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Paolo Bonzini]

On 25/07/2018 14:50, Vitaly Kuznetsov wrote:
>>
>> But is L0 allowed to write to hv_clean_fields?  
> It is kinda expected to: currently I reset it in vmx_vcpu_run() and (if
> I remember correctly) L1 Hyper-V only clears bits in this mask when it
> touches certain fields so if we don't set it to 'all clean' it stays
> zeroed forever.

Oh, good.  I didn't understand it was bidirectional.

 So nothing stops us from doing
> 
>        if (hv_evmcs && vmx->nested.dirty_vmcs12)
>                hv_evmcs->hv_clean_fields &=
>                        ~HV_VMX_ENLIGHTENED_CLEAN_FIELD_ALL;
> 
> in prepare_vmcs02() I guess.

In prepare_vmcs02, or rather in the enlightened VMPTRLD?

>> One possibility is to
>> add a dirty_evmcs field to struct nested_vmx, and "OR" ~hv_clean_fields
>> into it at the beginning of prepare_vmcs02.
>>
>> Something like
>>
>> 	if (vmx->nested.hv_evmcs) {
>> 		vmx->nested.dirty_evmcs |=
>> 			~vmx->nested.hv_evmcs->hv_clean_fields;
>> 		prepare_vmcs02_full(vcpu, vmcs12,
>> 				    vmx->nested.dirty_evmcs);
>> 	} else if (vmx->nested.dirty_vmcs12) {
>> 		prepare_vmcs02_full(vcpu, vmcs12, ~0);
>> 	}
>>
>> 	...
>> 	vmx->nested.dirty_evmcs = 0;
>> 	vmx->nested.dirty_vmcs12 = false;
>>
>> ?
>>
> I think we can even get away with a local variable in prepare_vmcs02()
> and pass it to prepare_vmcs02_full(), no need to have it in struct
> nested_vmx. But I would slightly prefer to just reset
> hv_evmcs->hv_clean_fields when vmcs12 is dirty.

Yes, that's even better.

Paolo

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 25/07/2018 14:50, Vitaly Kuznetsov wrote:
>>>
>>> But is L0 allowed to write to hv_clean_fields?  
>> It is kinda expected to: currently I reset it in vmx_vcpu_run() and (if
>> I remember correctly) L1 Hyper-V only clears bits in this mask when it
>> touches certain fields so if we don't set it to 'all clean' it stays
>> zeroed forever.
>
> Oh, good.  I didn't understand it was bidirectional.
>
>  So nothing stops us from doing
>> 
>>        if (hv_evmcs && vmx->nested.dirty_vmcs12)
>>                hv_evmcs->hv_clean_fields &=
>>                        ~HV_VMX_ENLIGHTENED_CLEAN_FIELD_ALL;
>> 
>> in prepare_vmcs02() I guess.
>
> In prepare_vmcs02, or rather in the enlightened VMPTRLD?
>

Doing it in nested_vmx_handle_enlightened_vmptrld() is even better: we
can simplify copy_enlightened_to_vmcs12() too!

The other place where we set dirty_vmcs12 is the newly introduced
vmx_set_nested_state() but I think I'm going to add support for eVMCS
there later and just return something like -ENOTSUPP for now. Too many
people work on nested simultaneously :-)

-- 
  Vitaly

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Paolo Bonzini]

On 25/07/2018 15:26, Vitaly Kuznetsov wrote:
> Doing it in nested_vmx_handle_enlightened_vmptrld() is even better: we
> can simplify copy_enlightened_to_vmcs12() too!

Yes, that's true!

> The other place where we set dirty_vmcs12 is the newly introduced
> vmx_set_nested_state() but I think I'm going to add support for eVMCS
> there later and just return something like -ENOTSUPP for now. Too many
> people work on nested simultaneously :-)

Hmm, I think that means you have to put the clean fields in the
vmx_nested struct.  Then it's really easy in vmx_set_nested_state to
clear all the clean bits.  Touching memory in vmx_set_nested_state is...
*puts on sunglasses* a touchy subject (see comment in kvm/queue's
enter_vmx_non_root_mode).

Thanks,

Paolo

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 25/07/2018 15:26, Vitaly Kuznetsov wrote:
>
>> The other place where we set dirty_vmcs12 is the newly introduced
>> vmx_set_nested_state() but I think I'm going to add support for eVMCS
>> there later and just return something like -ENOTSUPP for now. Too many
>> people work on nested simultaneously :-)
>
> Hmm, I think that means you have to put the clean fields in the
> vmx_nested struct.  Then it's really easy in vmx_set_nested_state to
> clear all the clean bits.  Touching memory in vmx_set_nested_state is...
> *puts on sunglasses* a touchy subject (see comment in kvm/queue's
> enter_vmx_non_root_mode).
>

Not only clean fields, in case we can't touch memory in
vmx_set_nested_state I guess we'll need to cache the whole eVMCS, just
like we save cached_shadow_vmcs12. Anyway, let's eat the elephant one
bite at a time :-)

-- 
  Vitaly

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Paolo Bonzini]

On 25/07/2018 16:13, Vitaly Kuznetsov wrote:
> Paolo Bonzini <pbonzini@redhat.com> writes:
> 
>> On 25/07/2018 15:26, Vitaly Kuznetsov wrote:
>>
>>> The other place where we set dirty_vmcs12 is the newly introduced
>>> vmx_set_nested_state() but I think I'm going to add support for eVMCS
>>> there later and just return something like -ENOTSUPP for now. Too many
>>> people work on nested simultaneously :-)
>>
>> Hmm, I think that means you have to put the clean fields in the
>> vmx_nested struct.  Then it's really easy in vmx_set_nested_state to
>> clear all the clean bits.  Touching memory in vmx_set_nested_state is...
>> *puts on sunglasses* a touchy subject (see comment in kvm/queue's
>> enter_vmx_non_root_mode).
>>
> 
> Not only clean fields, in case we can't touch memory in
> vmx_set_nested_state I guess we'll need to cache the whole eVMCS, just
> like we save cached_shadow_vmcs12. Anyway, let's eat the elephant one
> bite at a time :-)

You're right, but do we actually need to save the vmcs12 if eVMCS is
active?  The format would be different, but the size wouldn't.

Paolo

Re: [PATCH v2 6/6] KVM: nVMX: optimize prepare_vmcs02{,_full} for Enlightened VMCS case

[Vitaly Kuznetsov]

Paolo Bonzini <pbonzini@redhat.com> writes:

> On 25/07/2018 16:13, Vitaly Kuznetsov wrote:
>> Paolo Bonzini <pbonzini@redhat.com> writes:
>> 
>>> On 25/07/2018 15:26, Vitaly Kuznetsov wrote:
>>>
>>>> The other place where we set dirty_vmcs12 is the newly introduced
>>>> vmx_set_nested_state() but I think I'm going to add support for eVMCS
>>>> there later and just return something like -ENOTSUPP for now. Too many
>>>> people work on nested simultaneously :-)
>>>
>>> Hmm, I think that means you have to put the clean fields in the
>>> vmx_nested struct.  Then it's really easy in vmx_set_nested_state to
>>> clear all the clean bits.  Touching memory in vmx_set_nested_state is...
>>> *puts on sunglasses* a touchy subject (see comment in kvm/queue's
>>> enter_vmx_non_root_mode).
>>>
>> 
>> Not only clean fields, in case we can't touch memory in
>> vmx_set_nested_state I guess we'll need to cache the whole eVMCS, just
>> like we save cached_shadow_vmcs12. Anyway, let's eat the elephant one
>> bite at a time :-)
>
> You're right, but do we actually need to save the vmcs12 if eVMCS is
> active?  The format would be different, but the size wouldn't.

I think we don't: vmcs12 can always be re-constructed from eVMCS and
vice versa. The size is not exactly the same as we have some stuff
which is missing in eVMCS but we know that eVMCS will always fit into
vmcs12 area.

-- 
  Vitaly

Re: [PATCH v2 0/6] KVM: nVMX: Enlightened VMCS for Hyper-V on KVM

[Vitaly Kuznetsov]

Vitaly Kuznetsov <vkuznets@redhat.com> writes:

> Changes since v1 [Liran Alon]:
> - Split PATCH2 into two.
> - Add a comment explaining eVMCS versioning.
> - Convert evmcs_needs_write() to inline function.
> - Drop 'unlikely' in vmx_vcpu_run().
> - Add Reviewed-by tags.
>
> Description:
>
> This is an initial implementation of Enlightened VMCS for nested Hyper-V on
> KVM. Using it helps to spare 1500 cpu cycles for nested vmexit (tight cpuid
> loop in WS2016 with Hyper-V role on KVM: 15200 cycles -> 13700
> cycles).

It's been a month, ping :-)

-- 
  Vitaly