[PATCH v2 0/8] KVM: x86: enable cflushopt/clwb/pcommit and simplify code

[PATCH v2 0/8] KVM: x86: enable cflushopt/clwb/pcommit and simplify code

[Xiao Guangrong]

Changelog:
Thanks for Paolo's review, there are the changes in v2:
- use WARN_ON(1) instead of BUG() if PCOMMIT-exit happend for L1 guest
- drop set_clear_2nd_exec_ctrl() and use vmcs_{set,clear}_bits  instead
- improve commit log and adjust code style

This pachset enables clfushopt, clwb and pcommit instructions for guest which
are used by NVDIMM.

The specification locates at:
https://software.intel.com/sites/default/files/managed/0d/53/319433-022.pdf

patch 1 and patch 2 enable these three instructions for guest and other patches
simplify current VMX code

Xiao Guangrong (8):
  KVM: x86: allow guest to use cflushopt and clwb
  KVM: x86: add pcommit support
  KVM: VMX: drop rdtscp_enabled check in prepare_vmcs02()
  KVM: VMX: simplify rdtscp handling in vmx_cpuid_update()
  KVM: VMX: simplify invpcid handling in vmx_cpuid_update()
  KVM: VMX: unify SECONDARY_VM_EXEC_CONTROL update
  KVM: VMX: clean up bit operation on SECONDARY_VM_EXEC_CONTROL
  KVM: VMX: drop rdtscp_enabled field

 arch/x86/include/asm/vmx.h      |   2 +-
 arch/x86/include/uapi/asm/vmx.h |   4 +-
 arch/x86/kvm/cpuid.c            |   2 +-
 arch/x86/kvm/cpuid.h            |  16 +++++++
 arch/x86/kvm/vmx.c              | 103 ++++++++++++++++++----------------------
 5 files changed, 67 insertions(+), 60 deletions(-)

-- 
2.4.3

[PATCH v2 1/8] KVM: x86: allow guest to use cflushopt and clwb

[Xiao Guangrong]

Pass these CPU features to guest to enable them in guest

They are needed by nvdimm drivers

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/cpuid.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 2fbea25..962fc7d 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -348,7 +348,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) |
 		F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) |
 		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
-		F(AVX512CD);
+		F(AVX512CD) | F(CLFLUSHOPT) | F(CLWB);
 
 	/* cpuid 0xD.1.eax */
 	const u32 kvm_supported_word10_x86_features =
-- 
2.4.3

[PATCH v2 2/8] KVM: x86: add pcommit support

[Xiao Guangrong]

Pass PCOMMIT CPU feature to guest to enable PCOMMIT instruction

Currently we do not catch pcommit instruction for L1 guest and
allow L1 to catch this instruction for L2 if, as required by the spec,
L1 can enumerate the PCOMMIT instruction via CPUID:
| IA32_VMX_PROCBASED_CTLS2[53] (which enumerates support for the
| 1-setting of PCOMMIT exiting) is always the same as
| CPUID.07H:EBX.PCOMMIT[bit 22]. Thus, software can set PCOMMIT exiting
| to 1 if and only if the PCOMMIT instruction is enumerated via CPUID

The spec can be found at
https://software.intel.com/sites/default/files/managed/0d/53/319433-022.pdf

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/include/asm/vmx.h      |  2 +-
 arch/x86/include/uapi/asm/vmx.h |  4 +++-
 arch/x86/kvm/cpuid.c            |  2 +-
 arch/x86/kvm/cpuid.h            |  8 ++++++++
 arch/x86/kvm/vmx.c              | 29 ++++++++++++++++++++++++-----
 5 files changed, 37 insertions(+), 8 deletions(-)

diff --git a/arch/x86/include/asm/vmx.h b/arch/x86/include/asm/vmx.h
index 448b7ca..d25f32a 100644
--- a/arch/x86/include/asm/vmx.h
+++ b/arch/x86/include/asm/vmx.h
@@ -72,7 +72,7 @@
 #define SECONDARY_EXEC_SHADOW_VMCS              0x00004000
 #define SECONDARY_EXEC_ENABLE_PML               0x00020000
 #define SECONDARY_EXEC_XSAVES			0x00100000
-
+#define SECONDARY_EXEC_PCOMMIT			0x00200000
 
 #define PIN_BASED_EXT_INTR_MASK                 0x00000001
 #define PIN_BASED_NMI_EXITING                   0x00000008
diff --git a/arch/x86/include/uapi/asm/vmx.h b/arch/x86/include/uapi/asm/vmx.h
index 37fee27..5b15d94 100644
--- a/arch/x86/include/uapi/asm/vmx.h
+++ b/arch/x86/include/uapi/asm/vmx.h
@@ -78,6 +78,7 @@
 #define EXIT_REASON_PML_FULL            62
 #define EXIT_REASON_XSAVES              63
 #define EXIT_REASON_XRSTORS             64
+#define EXIT_REASON_PCOMMIT             65
 
 #define VMX_EXIT_REASONS \
 	{ EXIT_REASON_EXCEPTION_NMI,         "EXCEPTION_NMI" }, \
@@ -126,7 +127,8 @@
 	{ EXIT_REASON_INVVPID,               "INVVPID" }, \
 	{ EXIT_REASON_INVPCID,               "INVPCID" }, \
 	{ EXIT_REASON_XSAVES,                "XSAVES" }, \
-	{ EXIT_REASON_XRSTORS,               "XRSTORS" }
+	{ EXIT_REASON_XRSTORS,               "XRSTORS" }, \
+	{ EXIT_REASON_PCOMMIT,               "PCOMMIT" }
 
 #define VMX_ABORT_SAVE_GUEST_MSR_FAIL        1
 #define VMX_ABORT_LOAD_HOST_MSR_FAIL         4
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 962fc7d..faeb0b3 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -348,7 +348,7 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 		F(FSGSBASE) | F(BMI1) | F(HLE) | F(AVX2) | F(SMEP) |
 		F(BMI2) | F(ERMS) | f_invpcid | F(RTM) | f_mpx | F(RDSEED) |
 		F(ADX) | F(SMAP) | F(AVX512F) | F(AVX512PF) | F(AVX512ER) |
-		F(AVX512CD) | F(CLFLUSHOPT) | F(CLWB);
+		F(AVX512CD) | F(CLFLUSHOPT) | F(CLWB) | F(PCOMMIT);
 
 	/* cpuid 0xD.1.eax */
 	const u32 kvm_supported_word10_x86_features =
diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h
index dd05b9c..aed7bfe 100644
--- a/arch/x86/kvm/cpuid.h
+++ b/arch/x86/kvm/cpuid.h
@@ -133,4 +133,12 @@ static inline bool guest_cpuid_has_mpx(struct kvm_vcpu *vcpu)
 	best = kvm_find_cpuid_entry(vcpu, 7, 0);
 	return best && (best->ebx & bit(X86_FEATURE_MPX));
 }
+
+static inline bool guest_cpuid_has_pcommit(struct kvm_vcpu *vcpu)
+{
+	struct kvm_cpuid_entry2 *best;
+
+	best = kvm_find_cpuid_entry(vcpu, 7, 0);
+	return best && (best->ebx & bit(X86_FEATURE_PCOMMIT));
+}
 #endif
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index da1590e..073cbc8 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -2474,7 +2474,8 @@ static void nested_vmx_setup_ctls_msrs(struct vcpu_vmx *vmx)
 		SECONDARY_EXEC_APIC_REGISTER_VIRT |
 		SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
 		SECONDARY_EXEC_WBINVD_EXITING |
-		SECONDARY_EXEC_XSAVES;
+		SECONDARY_EXEC_XSAVES |
+		SECONDARY_EXEC_PCOMMIT;
 
 	if (enable_ept) {
 		/* nested EPT: emulate EPT also to L1 */
@@ -3015,7 +3016,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
 			SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
 			SECONDARY_EXEC_SHADOW_VMCS |
 			SECONDARY_EXEC_XSAVES |
-			SECONDARY_EXEC_ENABLE_PML;
+			SECONDARY_EXEC_ENABLE_PML |
+			SECONDARY_EXEC_PCOMMIT;
 		if (adjust_vmx_controls(min2, opt2,
 					MSR_IA32_VMX_PROCBASED_CTLS2,
 					&_cpu_based_2nd_exec_control) < 0)
@@ -4570,6 +4572,9 @@ static u32 vmx_secondary_exec_control(struct vcpu_vmx *vmx)
 	/* PML is enabled/disabled in creating/destorying vcpu */
 	exec_control &= ~SECONDARY_EXEC_ENABLE_PML;
 
+	/* Currently, we allow L1 guest to directly run pcommit instruction. */
+	exec_control &= ~SECONDARY_EXEC_PCOMMIT;
+
 	return exec_control;
 }
 
@@ -4613,10 +4618,9 @@ static int vmx_vcpu_setup(struct vcpu_vmx *vmx)
 
 	vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, vmx_exec_control(vmx));
 
-	if (cpu_has_secondary_exec_ctrls()) {
+	if (cpu_has_secondary_exec_ctrls())
 		vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
 				vmx_secondary_exec_control(vmx));
-	}
 
 	if (vmx_vm_has_apicv(vmx->vcpu.kvm)) {
 		vmcs_write64(EOI_EXIT_BITMAP0, 0);
@@ -7219,6 +7223,13 @@ static int handle_pml_full(struct kvm_vcpu *vcpu)
 	return 1;
 }
 
+static int handle_pcommit(struct kvm_vcpu *vcpu)
+{
+	/* we never catch pcommit instruct for L1 guest. */
+	WARN_ON(1);
+	return 1;
+}
+
 /*
  * The exit handlers return 1 if the exit was handled fully and guest execution
  * may resume.  Otherwise they set the kvm_run parameter to indicate what needs
@@ -7269,6 +7280,7 @@ static int (*const kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu) = {
 	[EXIT_REASON_XSAVES]                  = handle_xsaves,
 	[EXIT_REASON_XRSTORS]                 = handle_xrstors,
 	[EXIT_REASON_PML_FULL]		      = handle_pml_full,
+	[EXIT_REASON_PCOMMIT]                 = handle_pcommit,
 };
 
 static const int kvm_vmx_max_exit_handlers =
@@ -7570,6 +7582,8 @@ static bool nested_vmx_exit_handled(struct kvm_vcpu *vcpu)
 		 * the XSS exit bitmap in vmcs12.
 		 */
 		return nested_cpu_has2(vmcs12, SECONDARY_EXEC_XSAVES);
+	case EXIT_REASON_PCOMMIT:
+		return nested_cpu_has2(vmcs12, SECONDARY_EXEC_PCOMMIT);
 	default:
 		return true;
 	}
@@ -8698,6 +8712,10 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 		if (best)
 			best->ebx &= ~bit(X86_FEATURE_INVPCID);
 	}
+
+	if (!guest_cpuid_has_pcommit(vcpu) && nested)
+		vmx->nested.nested_vmx_secondary_ctls_high &=
+			~SECONDARY_EXEC_PCOMMIT;
 }
 
 static void vmx_set_supported_cpuid(u32 func, struct kvm_cpuid_entry2 *entry)
@@ -9311,7 +9329,8 @@ static void prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 		exec_control &= ~(SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |
 				  SECONDARY_EXEC_RDTSCP |
 				  SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY |
-				  SECONDARY_EXEC_APIC_REGISTER_VIRT);
+				  SECONDARY_EXEC_APIC_REGISTER_VIRT |
+				  SECONDARY_EXEC_PCOMMIT);
 		if (nested_cpu_has(vmcs12,
 				CPU_BASED_ACTIVATE_SECONDARY_CONTROLS))
 			exec_control |= vmcs12->secondary_vm_exec_control;
-- 
2.4.3

[PATCH v2 3/8] KVM: VMX: drop rdtscp_enabled check in prepare_vmcs02()

[Xiao Guangrong]

SECONDARY_EXEC_RDTSCP set for L2 guest comes from vmcs12

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/vmx.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 073cbc8..61d44b0 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -9323,8 +9323,7 @@ static void prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12)
 
 	if (cpu_has_secondary_exec_ctrls()) {
 		exec_control = vmx_secondary_exec_control(vmx);
-		if (!vmx->rdtscp_enabled)
-			exec_control &= ~SECONDARY_EXEC_RDTSCP;
+
 		/* Take the following fields only from vmcs12 */
 		exec_control &= ~(SECONDARY_EXEC_VIRTUALIZE_APIC_ACCESSES |
 				  SECONDARY_EXEC_RDTSCP |
-- 
2.4.3

[PATCH v2 4/8] KVM: VMX: simplify rdtscp handling in vmx_cpuid_update()

[Xiao Guangrong]

if vmx_rdtscp_supported() is true SECONDARY_EXEC_RDTSCP must
have already been set in current vmcs by
vmx_secondary_exec_control()

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/vmx.c | 17 ++++++++---------
 1 file changed, 8 insertions(+), 9 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 61d44b0..bcc69de 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -8678,16 +8678,15 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 	vmx->rdtscp_enabled = false;
 	if (vmx_rdtscp_supported()) {
 		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-		if (exec_control & SECONDARY_EXEC_RDTSCP) {
-			best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
-			if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
-				vmx->rdtscp_enabled = true;
-			else {
-				exec_control &= ~SECONDARY_EXEC_RDTSCP;
-				vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
-						exec_control);
-			}
+		best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
+		if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
+			vmx->rdtscp_enabled = true;
+		else {
+			exec_control &= ~SECONDARY_EXEC_RDTSCP;
+			vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
+					exec_control);
 		}
+
 		if (nested && !vmx->rdtscp_enabled)
 			vmx->nested.nested_vmx_secondary_ctls_high &=
 				~SECONDARY_EXEC_RDTSCP;
-- 
2.4.3

[PATCH v2 5/8] KVM: VMX: simplify invpcid handling in vmx_cpuid_update()

[Xiao Guangrong]

If vmx_invpcid_supported() is true, second execution control
filed must be supported and SECONDARY_EXEC_ENABLE_INVPCID
must have already been set in current vmcs by
vmx_secondary_exec_control()

If vmx_invpcid_supported() is false, no need to clear
SECONDARY_EXEC_ENABLE_INVPCID

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/vmx.c | 17 +++++------------
 1 file changed, 5 insertions(+), 12 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index bcc69de..97e3340 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -8695,19 +8695,12 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 	/* Exposing INVPCID only when PCID is exposed */
 	best = kvm_find_cpuid_entry(vcpu, 0x7, 0);
 	if (vmx_invpcid_supported() &&
-	    best && (best->ebx & bit(X86_FEATURE_INVPCID)) &&
-	    guest_cpuid_has_pcid(vcpu)) {
+	    (!best || !(best->ebx & bit(X86_FEATURE_INVPCID)) ||
+	    !guest_cpuid_has_pcid(vcpu))) {
 		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-		exec_control |= SECONDARY_EXEC_ENABLE_INVPCID;
-		vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
-			     exec_control);
-	} else {
-		if (cpu_has_secondary_exec_ctrls()) {
-			exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-			exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID;
-			vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
-				     exec_control);
-		}
+		exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID;
+		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+
 		if (best)
 			best->ebx &= ~bit(X86_FEATURE_INVPCID);
 	}
-- 
2.4.3

[PATCH v2 6/8] KVM: VMX: unify SECONDARY_VM_EXEC_CONTROL update

[Xiao Guangrong]

Unify the update in vmx_cpuid_update()

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/vmx.c | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 97e3340..5a074d0 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -8673,19 +8673,15 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 {
 	struct kvm_cpuid_entry2 *best;
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
-	u32 exec_control;
+	u32 clear_exe_ctrl = 0;
 
 	vmx->rdtscp_enabled = false;
 	if (vmx_rdtscp_supported()) {
-		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
 		best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
 		if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
 			vmx->rdtscp_enabled = true;
-		else {
-			exec_control &= ~SECONDARY_EXEC_RDTSCP;
-			vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
-					exec_control);
-		}
+		else
+			clear_exe_ctrl |= SECONDARY_EXEC_RDTSCP;
 
 		if (nested && !vmx->rdtscp_enabled)
 			vmx->nested.nested_vmx_secondary_ctls_high &=
@@ -8697,14 +8693,19 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 	if (vmx_invpcid_supported() &&
 	    (!best || !(best->ebx & bit(X86_FEATURE_INVPCID)) ||
 	    !guest_cpuid_has_pcid(vcpu))) {
-		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-		exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID;
-		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+		clear_exe_ctrl |= SECONDARY_EXEC_ENABLE_INVPCID;
 
 		if (best)
 			best->ebx &= ~bit(X86_FEATURE_INVPCID);
 	}
 
+	if (clear_exe_ctrl) {
+		u32 exec_ctl = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
+
+		exec_ctl &= ~clear_exe_ctrl;
+		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_ctl);
+	}
+
 	if (!guest_cpuid_has_pcommit(vcpu) && nested)
 		vmx->nested.nested_vmx_secondary_ctls_high &=
 			~SECONDARY_EXEC_PCOMMIT;
-- 
2.4.3

[PATCH v2 7/8] KVM: VMX: clean up bit operation on SECONDARY_VM_EXEC_CONTROL

[Xiao Guangrong]

Use vmcs_set_bits() and vmcs_clear_bits() to clean up the code

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/vmx.c | 31 ++++++++-----------------------
 1 file changed, 8 insertions(+), 23 deletions(-)

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 5a074d0..f18f744 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -6643,7 +6643,6 @@ static int nested_vmx_check_permission(struct kvm_vcpu *vcpu)
 
 static inline void nested_release_vmcs12(struct vcpu_vmx *vmx)
 {
-	u32 exec_control;
 	if (vmx->nested.current_vmptr == -1ull)
 		return;
 
@@ -6656,9 +6655,8 @@ static inline void nested_release_vmcs12(struct vcpu_vmx *vmx)
 		   they were modified */
 		copy_shadow_to_vmcs12(vmx);
 		vmx->nested.sync_shadow_vmcs = false;
-		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-		exec_control &= ~SECONDARY_EXEC_SHADOW_VMCS;
-		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+		vmcs_clear_bits(SECONDARY_VM_EXEC_CONTROL,
+				SECONDARY_EXEC_SHADOW_VMCS);
 		vmcs_write64(VMCS_LINK_POINTER, -1ull);
 	}
 	vmx->nested.posted_intr_nv = -1;
@@ -7054,7 +7052,6 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
 {
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	gpa_t vmptr;
-	u32 exec_control;
 
 	if (!nested_vmx_check_permission(vcpu))
 		return 1;
@@ -7086,9 +7083,8 @@ static int handle_vmptrld(struct kvm_vcpu *vcpu)
 		vmx->nested.current_vmcs12 = new_vmcs12;
 		vmx->nested.current_vmcs12_page = page;
 		if (enable_shadow_vmcs) {
-			exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-			exec_control |= SECONDARY_EXEC_SHADOW_VMCS;
-			vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+			vmcs_set_bits(SECONDARY_VM_EXEC_CONTROL,
+				      SECONDARY_EXEC_SHADOW_VMCS);
 			vmcs_write64(VMCS_LINK_POINTER,
 				     __pa(vmx->nested.current_shadow_vmcs));
 			vmx->nested.sync_shadow_vmcs = true;
@@ -7598,7 +7594,6 @@ static void vmx_get_exit_info(struct kvm_vcpu *vcpu, u64 *info1, u64 *info2)
 static int vmx_enable_pml(struct vcpu_vmx *vmx)
 {
 	struct page *pml_pg;
-	u32 exec_control;
 
 	pml_pg = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (!pml_pg)
@@ -7609,24 +7604,18 @@ static int vmx_enable_pml(struct vcpu_vmx *vmx)
 	vmcs_write64(PML_ADDRESS, page_to_phys(vmx->pml_pg));
 	vmcs_write16(GUEST_PML_INDEX, PML_ENTITY_NUM - 1);
 
-	exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-	exec_control |= SECONDARY_EXEC_ENABLE_PML;
-	vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+	vmcs_set_bits(SECONDARY_VM_EXEC_CONTROL, SECONDARY_EXEC_ENABLE_PML);
 
 	return 0;
 }
 
 static void vmx_disable_pml(struct vcpu_vmx *vmx)
 {
-	u32 exec_control;
-
 	ASSERT(vmx->pml_pg);
 	__free_page(vmx->pml_pg);
 	vmx->pml_pg = NULL;
 
-	exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-	exec_control &= ~SECONDARY_EXEC_ENABLE_PML;
-	vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
+	vmcs_clear_bits(SECONDARY_VM_EXEC_CONTROL, SECONDARY_EXEC_ENABLE_PML);
 }
 
 static void vmx_flush_pml_buffer(struct kvm_vcpu *vcpu)
@@ -8699,12 +8688,8 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 			best->ebx &= ~bit(X86_FEATURE_INVPCID);
 	}
 
-	if (clear_exe_ctrl) {
-		u32 exec_ctl = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-
-		exec_ctl &= ~clear_exe_ctrl;
-		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_ctl);
-	}
+	if (clear_exe_ctrl)
+		vmcs_clear_bits(SECONDARY_VM_EXEC_CONTROL, clear_exe_ctrl);
 
 	if (!guest_cpuid_has_pcommit(vcpu) && nested)
 		vmx->nested.nested_vmx_secondary_ctls_high &=
-- 
2.4.3

[PATCH v2 8/8] KVM: VMX: drop rdtscp_enabled field

[Xiao Guangrong]

Check cpuid bit instead of it

Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
---
 arch/x86/kvm/cpuid.h |  8 ++++++++
 arch/x86/kvm/vmx.c   | 19 ++++++-------------
 2 files changed, 14 insertions(+), 13 deletions(-)

diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h
index aed7bfe..d434ee9 100644
--- a/arch/x86/kvm/cpuid.h
+++ b/arch/x86/kvm/cpuid.h
@@ -141,4 +141,12 @@ static inline bool guest_cpuid_has_pcommit(struct kvm_vcpu *vcpu)
 	best = kvm_find_cpuid_entry(vcpu, 7, 0);
 	return best && (best->ebx & bit(X86_FEATURE_PCOMMIT));
 }
+
+static inline bool guest_cpuid_has_rdtscp(struct kvm_vcpu *vcpu)
+{
+	struct kvm_cpuid_entry2 *best;
+
+	best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
+	return best && (best->edx & bit(X86_FEATURE_RDTSCP));
+}
 #endif
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index f18f744..2e98e6d 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -532,8 +532,6 @@ struct vcpu_vmx {
 	s64 vnmi_blocked_time;
 	u32 exit_reason;
 
-	bool rdtscp_enabled;
-
 	/* Posted interrupt descriptor */
 	struct pi_desc pi_desc;
 
@@ -2207,7 +2205,7 @@ static void setup_msrs(struct vcpu_vmx *vmx)
 		if (index >= 0)
 			move_msr_up(vmx, index, save_nmsrs++);
 		index = __find_msr_index(vmx, MSR_TSC_AUX);
-		if (index >= 0 && vmx->rdtscp_enabled)
+		if (index >= 0 && guest_cpuid_has_rdtscp(&vmx->vcpu))
 			move_msr_up(vmx, index, save_nmsrs++);
 		/*
 		 * MSR_STAR is only needed on long mode guests, and only
@@ -2674,7 +2672,7 @@ static int vmx_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		msr_info->data = vcpu->arch.ia32_xss;
 		break;
 	case MSR_TSC_AUX:
-		if (!to_vmx(vcpu)->rdtscp_enabled)
+		if (!guest_cpuid_has_rdtscp(vcpu))
 			return 1;
 		/* Otherwise falls through */
 	default:
@@ -2780,7 +2778,7 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 			clear_atomic_switch_msr(vmx, MSR_IA32_XSS);
 		break;
 	case MSR_TSC_AUX:
-		if (!vmx->rdtscp_enabled)
+		if (!guest_cpuid_has_rdtscp(vcpu))
 			return 1;
 		/* Check reserved bit, higher 32 bits should be zero */
 		if ((data >> 32) != 0)
@@ -8664,15 +8662,10 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
 	u32 clear_exe_ctrl = 0;
 
-	vmx->rdtscp_enabled = false;
-	if (vmx_rdtscp_supported()) {
-		best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
-		if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
-			vmx->rdtscp_enabled = true;
-		else
-			clear_exe_ctrl |= SECONDARY_EXEC_RDTSCP;
+	if (vmx_rdtscp_supported() && !guest_cpuid_has_rdtscp(vcpu)) {
+		clear_exe_ctrl |= SECONDARY_EXEC_RDTSCP;
 
-		if (nested && !vmx->rdtscp_enabled)
+		if (nested)
 			vmx->nested.nested_vmx_secondary_ctls_high &=
 				~SECONDARY_EXEC_RDTSCP;
 	}
-- 
2.4.3

Re: [PATCH v2 2/8] KVM: x86: add pcommit support

[Paolo Bonzini]

On 09/09/2015 08:05, Xiao Guangrong wrote:
> +	if (!guest_cpuid_has_pcommit(vcpu) && nested)
> +		vmx->nested.nested_vmx_secondary_ctls_high &=
> +			~SECONDARY_EXEC_PCOMMIT;

It is legal to set CPUID multiple times, so I think we need

        if (static_cpu_has(X86_FEATURE_PCOMMIT) && nested) {
                if (guest_cpuid_has_pcommit(vcpu))
                        vmx->nested.nested_vmx_secondary_ctls_high |=
                                SECONDARY_EXEC_PCOMMIT;
                else
                        vmx->nested.nested_vmx_secondary_ctls_high &=
                                ~SECONDARY_EXEC_PCOMMIT;
        }

The INVPCID exit probably should be handled the same way as PCOMMIT;
moving both to a new nested_vmx_update_ctls_msrs probably makes sense.

This can be done on top of this series.

Paolo

Re: [PATCH v2 6/8] KVM: VMX: unify SECONDARY_VM_EXEC_CONTROL update

[Paolo Bonzini]

On 09/09/2015 08:05, Xiao Guangrong wrote:
> Unify the update in vmx_cpuid_update()
> 
> Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>

What if we instead start fresh from vmx_secondary_exec_control, like this:

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 4504cbc1b287..98c4d3f1653a 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -8660,11 +8660,26 @@ static int vmx_get_lpage_level(void)
 		return PT_PDPE_LEVEL;
 }
 
+static void vmcs_set_secondary_exec_control(u32 new_ctl)
+{
+	/*
+	 * These bits in the secondary execution controls field
+	 * are dynamic, the others are mostly based on the hypervisor
+	 * architecture and the guest's CPUID.  Do not touch the
+	 * dynamic bits.
+	 */
+	u32 mask = SECONDARY_EXEC_ENABLE_PML | SECONDARY_EXEC_SHADOW_VMCS;
+	u32 cur_ctl = vmcs_read32(SECONDARY_EXEC_CONTROL);
+
+	vmcs_write32(SECONDARY_EXEC_CONTROL,
+		     (new_ctl & ~mask) | (cur_ctl & mask));
+}
+
 static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 {
 	struct kvm_cpuid_entry2 *best;
 	struct vcpu_vmx *vmx = to_vmx(vcpu);
-	u32 clear_exe_ctrl = 0;
+	u32 secondary_exec_ctl = vmx_secondary_exec_control(vmx);
 
 	vmx->rdtscp_enabled = false;
 	if (vmx_rdtscp_supported()) {
@@ -8672,7 +8681,7 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 		if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
 			vmx->rdtscp_enabled = true;
 		else
-			clear_exe_ctrl |= SECONDARY_EXEC_RDTSCP;
+			secondary_exec_ctl &= ~SECONDARY_EXEC_RDTSCP;
 
 		if (nested) {
 			if (vmx->rdtscp_enabled)
@@ -8689,18 +8698,13 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
 	if (vmx_invpcid_supported() &&
 	    (!best || !(best->ebx & bit(X86_FEATURE_INVPCID)) ||
 	    !guest_cpuid_has_pcid(vcpu))) {
-		clear_exe_ctrl |= SECONDARY_EXEC_ENABLE_INVPCID;
+		secondary_exec_ctl &= ~SECONDARY_EXEC_ENABLE_INVPCID;
 
 		if (best)
 			best->ebx &= ~bit(X86_FEATURE_INVPCID);
 	}
 
-	if (clear_exe_ctrl) {
-		u32 exec_ctl = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
-
-		exec_ctl &= ~clear_exe_ctrl;
-		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_ctl);
-	}
+	vmcs_set_secondary_exec_control(secondary_exec_ctl);
 
 	if (static_cpu_has(X86_FEATURE_PCOMMIT) && nested) {
 		if (guest_cpuid_has_pcommit(vcpu))

> ---
>  arch/x86/kvm/vmx.c | 21 +++++++++++----------
>  1 file changed, 11 insertions(+), 10 deletions(-)
> 
> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
> index 97e3340..5a074d0 100644
> --- a/arch/x86/kvm/vmx.c
> +++ b/arch/x86/kvm/vmx.c
> @@ -8673,19 +8673,15 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
>  {
>  	struct kvm_cpuid_entry2 *best;
>  	struct vcpu_vmx *vmx = to_vmx(vcpu);
> -	u32 exec_control;
> +	u32 clear_exe_ctrl = 0;
>  
>  	vmx->rdtscp_enabled = false;
>  	if (vmx_rdtscp_supported()) {
> -		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
>  		best = kvm_find_cpuid_entry(vcpu, 0x80000001, 0);
>  		if (best && (best->edx & bit(X86_FEATURE_RDTSCP)))
>  			vmx->rdtscp_enabled = true;
> -		else {
> -			exec_control &= ~SECONDARY_EXEC_RDTSCP;
> -			vmcs_write32(SECONDARY_VM_EXEC_CONTROL,
> -					exec_control);
> -		}
> +		else
> +			clear_exe_ctrl |= SECONDARY_EXEC_RDTSCP;
>  
>  		if (nested && !vmx->rdtscp_enabled)
>  			vmx->nested.nested_vmx_secondary_ctls_high &=
> @@ -8697,14 +8693,19 @@ static void vmx_cpuid_update(struct kvm_vcpu *vcpu)
>  	if (vmx_invpcid_supported() &&
>  	    (!best || !(best->ebx & bit(X86_FEATURE_INVPCID)) ||
>  	    !guest_cpuid_has_pcid(vcpu))) {
> -		exec_control = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
> -		exec_control &= ~SECONDARY_EXEC_ENABLE_INVPCID;
> -		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_control);
> +		clear_exe_ctrl |= SECONDARY_EXEC_ENABLE_INVPCID;
>  
>  		if (best)
>  			best->ebx &= ~bit(X86_FEATURE_INVPCID);
>  	}
>  
> +	if (clear_exe_ctrl) {
> +		u32 exec_ctl = vmcs_read32(SECONDARY_VM_EXEC_CONTROL);
> +
> +		exec_ctl &= ~clear_exe_ctrl;
> +		vmcs_write32(SECONDARY_VM_EXEC_CONTROL, exec_ctl);
> +	}
> +
>  	if (!guest_cpuid_has_pcommit(vcpu) && nested)
>  		vmx->nested.nested_vmx_secondary_ctls_high &=
>  			~SECONDARY_EXEC_PCOMMIT;
>