KVM: MMU: prepopulate the shadow on invlpg

If the guest executes invlpg, peek into the pagetable and attempt to
prepopulate the shadow entry.

Also stop dirty fault updates from interfering with the fork detector.

2% improvement on RHEL3/AIM7.

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

4 files changed, 40 insertions, 15 deletions

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 65b1ed295698..97215a458e5f 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -602,7 +602,8 @@ unsigned long segment_base(u16 selector);
 
 void kvm_mmu_flush_tlb(struct kvm_vcpu *vcpu);
 void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
-		       const u8 *new, int bytes);
+		       const u8 *new, int bytes,
+		       bool guest_initiated);
 int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva);
 void __kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu);
 int kvm_mmu_load(struct kvm_vcpu *vcpu);
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index cbac9e4b156f..863baf70506e 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -2441,7 +2441,8 @@ static void kvm_mmu_access_page(struct kvm_vcpu *vcpu, gfn_t gfn)
 }
 
 void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
-		       const u8 *new, int bytes)
+		       const u8 *new, int bytes,
+		       bool guest_initiated)
 {
 	gfn_t gfn = gpa >> PAGE_SHIFT;
 	struct kvm_mmu_page *sp;
@@ -2467,15 +2468,17 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
 	kvm_mmu_free_some_pages(vcpu);
 	++vcpu->kvm->stat.mmu_pte_write;
 	kvm_mmu_audit(vcpu, "pre pte write");
-	if (gfn == vcpu->arch.last_pt_write_gfn
-	    && !last_updated_pte_accessed(vcpu)) {
-		++vcpu->arch.last_pt_write_count;
-		if (vcpu->arch.last_pt_write_count >= 3)
-			flooded = 1;
-	} else {
-		vcpu->arch.last_pt_write_gfn = gfn;
-		vcpu->arch.last_pt_write_count = 1;
-		vcpu->arch.last_pte_updated = NULL;
+	if (guest_initiated) {
+		if (gfn == vcpu->arch.last_pt_write_gfn
+		    && !last_updated_pte_accessed(vcpu)) {
+			++vcpu->arch.last_pt_write_count;
+			if (vcpu->arch.last_pt_write_count >= 3)
+				flooded = 1;
+		} else {
+			vcpu->arch.last_pt_write_gfn = gfn;
+			vcpu->arch.last_pt_write_count = 1;
+			vcpu->arch.last_pte_updated = NULL;
+		}
 	}
 	index = kvm_page_table_hashfn(gfn);
 	bucket = &vcpu->kvm->arch.mmu_page_hash[index];
@@ -2615,9 +2618,7 @@ 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;
 }
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index e644d81979b6..d20640154216 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -82,6 +82,7 @@ struct shadow_walker {
 	int *ptwrite;
 	pfn_t pfn;
 	u64 *sptep;
+	gpa_t pte_gpa;
 };
 
 static gfn_t gpte_to_gfn(pt_element_t gpte)
@@ -222,7 +223,7 @@ walk:
 		if (ret)
 			goto walk;
 		pte |= PT_DIRTY_MASK;
-		kvm_mmu_pte_write(vcpu, pte_gpa, (u8 *)&pte, sizeof(pte));
+		kvm_mmu_pte_write(vcpu, pte_gpa, (u8 *)&pte, sizeof(pte), 0);
 		walker->ptes[walker->level - 1] = pte;
 	}
 
@@ -468,8 +469,15 @@ static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw,
 				      struct kvm_vcpu *vcpu, u64 addr,
 				      u64 *sptep, int level)
 {
+	struct shadow_walker *sw =
+		container_of(_sw, struct shadow_walker, walker);
 
 	if (level == PT_PAGE_TABLE_LEVEL) {
+		struct kvm_mmu_page *sp = page_header(__pa(sptep));
+
+		sw->pte_gpa = (sp->gfn << PAGE_SHIFT);
+		sw->pte_gpa += (sptep - sp->spt) * sizeof(pt_element_t);
+
 		if (is_shadow_present_pte(*sptep))
 			rmap_remove(vcpu->kvm, sptep);
 		set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
@@ -482,11 +490,26 @@ static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw,
 
 static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva)
 {
+	pt_element_t gpte;
 	struct shadow_walker walker = {
 		.walker = { .entry = FNAME(shadow_invlpg_entry), },
+		.pte_gpa = -1,
 	};
 
+	spin_lock(&vcpu->kvm->mmu_lock);
 	walk_shadow(&walker.walker, vcpu, gva);
+	spin_unlock(&vcpu->kvm->mmu_lock);
+	if (walker.pte_gpa == -1)
+		return;
+	if (kvm_read_guest_atomic(vcpu->kvm, walker.pte_gpa, &gpte,
+				  sizeof(pt_element_t)))
+		return;
+	if (is_present_pte(gpte) && (gpte & PT_ACCESSED_MASK)) {
+		if (mmu_topup_memory_caches(vcpu))
+			return;
+		kvm_mmu_pte_write(vcpu, walker.pte_gpa, (const u8 *)&gpte,
+				  sizeof(pt_element_t), 0);
+	}
 }
 
 static gpa_t FNAME(gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t vaddr)
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 774db00d2db6..ba102879de33 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2046,7 +2046,7 @@ int emulator_write_phys(struct kvm_vcpu *vcpu, gpa_t gpa,
 	ret = kvm_write_guest(vcpu->kvm, gpa, val, bytes);
 	if (ret < 0)
 		return 0;
-	kvm_mmu_pte_write(vcpu, gpa, val, bytes);
+	kvm_mmu_pte_write(vcpu, gpa, val, bytes, 1);
 	return 1;
 }