KVM: x86: trap invlpg

With pages out of sync invlpg needs to be trapped. For now simply nuke
the entry.

Untested on AMD.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>

5 files changed, 71 insertions, 5 deletions

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 9d8c4bb..e89af1d 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -877,6 +877,10 @@ static int nonpaging_sync_page(struct kvm_vcpu *vcpu,
 	return 1;
 }
 
+static void nonpaging_invlpg(struct kvm_vcpu *vcpu, gva_t gva)
+{
+}
+
 static struct kvm_mmu_page *kvm_mmu_lookup_page(struct kvm *kvm, gfn_t gfn)
 {
 	unsigned index;
@@ -1589,6 +1593,7 @@ static int nonpaging_init_context(struct kvm_vcpu *vcpu)
 	context->free = nonpaging_free;
 	context->prefetch_page = nonpaging_prefetch_page;
 	context->sync_page = nonpaging_sync_page;
+	context->invlpg = nonpaging_invlpg;
 	context->root_level = 0;
 	context->shadow_root_level = PT32E_ROOT_LEVEL;
 	context->root_hpa = INVALID_PAGE;
@@ -1637,6 +1642,7 @@ static int paging64_init_context_common(struct kvm_vcpu *vcpu, int level)
 	context->gva_to_gpa = paging64_gva_to_gpa;
 	context->prefetch_page = paging64_prefetch_page;
 	context->sync_page = paging64_sync_page;
+	context->invlpg = paging64_invlpg;
 	context->free = paging_free;
 	context->root_level = level;
 	context->shadow_root_level = level;
@@ -1659,6 +1665,7 @@ static int paging32_init_context(struct kvm_vcpu *vcpu)
 	context->free = paging_free;
 	context->prefetch_page = paging32_prefetch_page;
 	context->sync_page = paging32_sync_page;
+	context->invlpg = paging32_invlpg;
 	context->root_level = PT32_ROOT_LEVEL;
 	context->shadow_root_level = PT32E_ROOT_LEVEL;
 	context->root_hpa = INVALID_PAGE;
@@ -1679,6 +1686,7 @@ static int init_kvm_tdp_mmu(struct kvm_vcpu *vcpu)
 	context->free = nonpaging_free;
 	context->prefetch_page = nonpaging_prefetch_page;
 	context->sync_page = nonpaging_sync_page;
+	context->invlpg = nonpaging_invlpg;
 	context->shadow_root_level = kvm_x86_ops->get_tdp_level();
 	context->root_hpa = INVALID_PAGE;
 
@@ -2071,6 +2079,16 @@ out:
 }
 EXPORT_SYMBOL_GPL(kvm_mmu_page_fault);
 
+void kvm_mmu_invlpg(struct kvm_vcpu *vcpu, gva_t gva)
+{
+	spin_lock(&vcpu->kvm->mmu_lock);
+	vcpu->arch.mmu.invlpg(vcpu, gva);
+	spin_unlock(&vcpu->kvm->mmu_lock);
+	kvm_mmu_flush_tlb(vcpu);
+	++vcpu->stat.invlpg;
+}
+EXPORT_SYMBOL_GPL(kvm_mmu_invlpg);
+
 void kvm_enable_tdp(void)
 {
 	tdp_enabled = true;
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index 776fb6d..dc169e8 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -461,6 +461,31 @@ out_unlock:
 	return 0;
 }
 
+static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw,
+				      struct kvm_vcpu *vcpu, u64 addr,
+				      u64 *sptep, int level)
+{
+
+	if (level == PT_PAGE_TABLE_LEVEL) {
+		if (is_shadow_present_pte(*sptep))
+			rmap_remove(vcpu->kvm, sptep);
+		set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
+		return 1;
+	}
+	if (!is_shadow_present_pte(*sptep))
+		return 1;
+	return 0;
+}
+
+static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva)
+{
+	struct shadow_walker walker = {
+		.walker = { .entry = FNAME(shadow_invlpg_entry), },
+	};
+
+	walk_shadow(&walker.walker, vcpu, gva);
+}
+
 static gpa_t FNAME(gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t vaddr)
 {
 	struct guest_walker walker;
diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index 9b54550..9c4ce65 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -525,6 +525,7 @@ static void init_vmcb(struct vcpu_svm *svm)
 				(1ULL << INTERCEPT_CPUID) |
 				(1ULL << INTERCEPT_INVD) |
 				(1ULL << INTERCEPT_HLT) |
+				(1ULL << INTERCEPT_INVLPG) |
 				(1ULL << INTERCEPT_INVLPGA) |
 				(1ULL << INTERCEPT_IOIO_PROT) |
 				(1ULL << INTERCEPT_MSR_PROT) |
@@ -589,7 +590,8 @@ static void init_vmcb(struct vcpu_svm *svm)
 	if (npt_enabled) {
 		/* Setup VMCB for Nested Paging */
 		control->nested_ctl = 1;
-		control->intercept &= ~(1ULL << INTERCEPT_TASK_SWITCH);
+		control->intercept &= ~((1ULL << INTERCEPT_TASK_SWITCH) |
+					(1ULL << INTERCEPT_INVLPG));
 		control->intercept_exceptions &= ~(1 << PF_VECTOR);
 		control->intercept_cr_read &= ~(INTERCEPT_CR0_MASK|
 						INTERCEPT_CR3_MASK);
@@ -1164,6 +1166,13 @@ static int cpuid_interception(struct vcpu_svm *svm, struct kvm_run *kvm_run)
 	return 1;
 }
 
+static int invlpg_interception(struct vcpu_svm *svm, struct kvm_run *kvm_run)
+{
+	if (emulate_instruction(&svm->vcpu, kvm_run, 0, 0, 0) != EMULATE_DONE)
+		pr_unimpl(&svm->vcpu, "%s: failed\n", __func__);
+	return 1;
+}
+
 static int emulate_on_interception(struct vcpu_svm *svm,
 				   struct kvm_run *kvm_run)
 {
@@ -1417,7 +1426,7 @@ static int (*svm_exit_handlers[])(struct vcpu_svm *svm,
 	[SVM_EXIT_CPUID]			= cpuid_interception,
 	[SVM_EXIT_INVD]                         = emulate_on_interception,
 	[SVM_EXIT_HLT]				= halt_interception,
-	[SVM_EXIT_INVLPG]			= emulate_on_interception,
+	[SVM_EXIT_INVLPG]			= invlpg_interception,
 	[SVM_EXIT_INVLPGA]			= invalid_op_interception,
 	[SVM_EXIT_IOIO] 		  	= io_interception,
 	[SVM_EXIT_MSR]				= msr_interception,
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 025bf40..4556cc3 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1130,7 +1130,8 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
 	      CPU_BASED_CR3_STORE_EXITING |
 	      CPU_BASED_USE_IO_BITMAPS |
 	      CPU_BASED_MOV_DR_EXITING |
-	      CPU_BASED_USE_TSC_OFFSETING;
+	      CPU_BASED_USE_TSC_OFFSETING |
+	      CPU_BASED_INVLPG_EXITING;
 	opt = CPU_BASED_TPR_SHADOW |
 	      CPU_BASED_USE_MSR_BITMAPS |
 	      CPU_BASED_ACTIVATE_SECONDARY_CONTROLS;
@@ -1159,9 +1160,11 @@ static __init int setup_vmcs_config(struct vmcs_config *vmcs_conf)
 		_cpu_based_exec_control &= ~CPU_BASED_TPR_SHADOW;
 #endif
 	if (_cpu_based_2nd_exec_control & SECONDARY_EXEC_ENABLE_EPT) {
-		/* CR3 accesses don't need to cause VM Exits when EPT enabled */
+		/* CR3 accesses and invlpg don't need to cause VM Exits when EPT
+		   enabled */
 		min &= ~(CPU_BASED_CR3_LOAD_EXITING |
-			 CPU_BASED_CR3_STORE_EXITING);
+			 CPU_BASED_CR3_STORE_EXITING |
+			 CPU_BASED_INVLPG_EXITING);
 		if (adjust_vmx_controls(min, opt, MSR_IA32_VMX_PROCBASED_CTLS,
 					&_cpu_based_exec_control) < 0)
 			return -EIO;
@@ -2790,6 +2793,15 @@ static int handle_vmcall(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
 	return 1;
 }
 
+static int handle_invlpg(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
+{
+	u64 exit_qualification = vmcs_read64(EXIT_QUALIFICATION);
+
+	kvm_mmu_invlpg(vcpu, exit_qualification);
+	skip_emulated_instruction(vcpu);
+	return 1;
+}
+
 static int handle_wbinvd(struct kvm_vcpu *vcpu, struct kvm_run *kvm_run)
 {
 	skip_emulated_instruction(vcpu);
@@ -2958,6 +2970,7 @@ static int (*kvm_vmx_exit_handlers[])(struct kvm_vcpu *vcpu,
 	[EXIT_REASON_MSR_WRITE]               = handle_wrmsr,
 	[EXIT_REASON_PENDING_INTERRUPT]       = handle_interrupt_window,
 	[EXIT_REASON_HLT]                     = handle_halt,
+	[EXIT_REASON_INVLPG]		      = handle_invlpg,
 	[EXIT_REASON_VMCALL]                  = handle_vmcall,
 	[EXIT_REASON_TPR_BELOW_THRESHOLD]     = handle_tpr_below_threshold,
 	[EXIT_REASON_APIC_ACCESS]             = handle_apic_access,
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 88e6d9a..efee85b 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2341,6 +2341,7 @@ static unsigned long get_segment_base(struct kvm_vcpu *vcpu, int seg)
 
 int emulate_invlpg(struct kvm_vcpu *vcpu, gva_t address)
 {
+	kvm_mmu_invlpg(vcpu, address);
 	return X86EMUL_CONTINUE;
 }