KVM: X86: Change kvm_sync_page() to return true when remote flush is needed

Currently kvm_sync_page() returns true when there is any present spte.
But the return value is ignored in the callers.

Changing kvm_sync_page() to return true when remote flush is needed and
changing mmu->sync_page() not to directly flush can combine and reduce
remote flush requests.
Signed-off-by: NLai Jiangshan <laijs@linux.alibaba.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210918005636.3675-7-jiangshanlai@gmail.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

arch/x86/kvm/mmu/mmu.c

@@ -1795,7 +1795,7 @@ static void mark_unsync(u64 *spte)
 static int nonpaging_sync_page(struct kvm_vcpu *vcpu,
 			       struct kvm_mmu_page *sp)
 {
-	return 0;
+	return -1;
 }
 
 #define KVM_PAGE_ARRAY_NR 16
@@ -1909,12 +1909,14 @@ static void kvm_mmu_commit_zap_page(struct kvm *kvm,
 static bool kvm_sync_page(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp,
 			 struct list_head *invalid_list)
 {
-	if (vcpu->arch.mmu->sync_page(vcpu, sp) == 0) {
+	int ret = vcpu->arch.mmu->sync_page(vcpu, sp);
+
+	if (ret < 0) {
 		kvm_mmu_prepare_zap_page(vcpu->kvm, sp, invalid_list);
 		return false;
 	}
 
-	return true;
+	return !!ret;
 }
 
 static bool kvm_mmu_remote_flush_or_zap(struct kvm *kvm,
@@ -2024,6 +2026,7 @@ static int mmu_sync_children(struct kvm_vcpu *vcpu,
 	struct mmu_page_path parents;
 	struct kvm_mmu_pages pages;
 	LIST_HEAD(invalid_list);
+	bool flush = false;
 
 	while (mmu_unsync_walk(parent, &pages)) {
 		bool protected = false;
@@ -2033,25 +2036,27 @@ static int mmu_sync_children(struct kvm_vcpu *vcpu,
 
 		if (protected) {
 			kvm_flush_remote_tlbs(vcpu->kvm);
+			flush = false;
 		}
 
 		for_each_sp(pages, sp, parents, i) {
 			kvm_unlink_unsync_page(vcpu->kvm, sp);
-			kvm_sync_page(vcpu, sp, &invalid_list);
+			flush |= kvm_sync_page(vcpu, sp, &invalid_list);
 			mmu_pages_clear_parents(&parents);
 		}
 		if (need_resched() || rwlock_needbreak(&vcpu->kvm->mmu_lock)) {
-			kvm_mmu_remote_flush_or_zap(vcpu->kvm, &invalid_list, false);
+			kvm_mmu_remote_flush_or_zap(vcpu->kvm, &invalid_list, flush);
 			if (!can_yield) {
 				kvm_make_request(KVM_REQ_MMU_SYNC, vcpu);
 				return -EINTR;
 			}
 
 			cond_resched_rwlock_write(&vcpu->kvm->mmu_lock);
+			flush = false;
 		}
 	}
 
-	kvm_mmu_remote_flush_or_zap(vcpu->kvm, &invalid_list, false);
+	kvm_mmu_remote_flush_or_zap(vcpu->kvm, &invalid_list, flush);
 	return 0;
 }
 
@@ -2135,6 +2140,7 @@ static struct kvm_mmu_page *kvm_mmu_get_page(struct kvm_vcpu *vcpu,
 				break;
 
 			WARN_ON(!list_empty(&invalid_list));
+			kvm_flush_remote_tlbs(vcpu->kvm);
 		}
 
 		__clear_sp_write_flooding_count(sp);
@@ -4191,7 +4197,7 @@ static unsigned long get_cr3(struct kvm_vcpu *vcpu)
 }
 
 static bool sync_mmio_spte(struct kvm_vcpu *vcpu, u64 *sptep, gfn_t gfn,
-			   unsigned int access, int *nr_present)
+			   unsigned int access)
 {
 	if (unlikely(is_mmio_spte(*sptep))) {
 		if (gfn != get_mmio_spte_gfn(*sptep)) {
@@ -4199,7 +4205,6 @@ static bool sync_mmio_spte(struct kvm_vcpu *vcpu, u64 *sptep, gfn_t gfn,
 			return true;
 		}
 
-		(*nr_present)++;
 		mark_mmio_spte(vcpu, sptep, gfn, access);
 		return true;
 	}

arch/x86/kvm/mmu/paging_tmpl.h

@@ -1066,11 +1066,16 @@ static gpa_t FNAME(gva_to_gpa_nested)(struct kvm_vcpu *vcpu, gpa_t vaddr,
  * Using the cached information from sp->gfns is safe because:
  * - The spte has a reference to the struct page, so the pfn for a given gfn
  *   can't change unless all sptes pointing to it are nuked first.
+ *
+ * Returns
+ * < 0: the sp should be zapped
+ *   0: the sp is synced and no tlb flushing is required
+ * > 0: the sp is synced and tlb flushing is required
  */
 static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 {
 	union kvm_mmu_page_role mmu_role = vcpu->arch.mmu->mmu_role.base;
-	int i, nr_present = 0;
+	int i;
 	bool host_writable;
 	gpa_t first_pte_gpa;
 	int set_spte_ret = 0;
@@ -1098,7 +1103,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 	 */
 	if (WARN_ON_ONCE(sp->role.direct ||
 			 (sp->role.word ^ mmu_role.word) & ~sync_role_ign.word))
-		return 0;
+		return -1;
 
 	first_pte_gpa = FNAME(get_level1_sp_gpa)(sp);
 
@@ -1115,7 +1120,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 
 		if (kvm_vcpu_read_guest_atomic(vcpu, pte_gpa, &gpte,
 					       sizeof(pt_element_t)))
-			return 0;
+			return -1;
 
 		if (FNAME(prefetch_invalid_gpte)(vcpu, sp, &sp->spt[i], gpte)) {
 			set_spte_ret |= SET_SPTE_NEED_REMOTE_TLB_FLUSH;
@@ -1127,8 +1132,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 		pte_access &= FNAME(gpte_access)(gpte);
 		FNAME(protect_clean_gpte)(vcpu->arch.mmu, &pte_access, gpte);
 
-		if (sync_mmio_spte(vcpu, &sp->spt[i], gfn, pte_access,
-		      &nr_present))
+		if (sync_mmio_spte(vcpu, &sp->spt[i], gfn, pte_access))
 			continue;
 
 		if (gfn != sp->gfns[i]) {
@@ -1137,8 +1141,6 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 			continue;
 		}
 
-		nr_present++;
-
 		host_writable = sp->spt[i] & shadow_host_writable_mask;
 
 		set_spte_ret |= set_spte(vcpu, &sp->spt[i],
@@ -1147,10 +1149,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 					 true, false, host_writable);
 	}
 
-	if (set_spte_ret & SET_SPTE_NEED_REMOTE_TLB_FLUSH)
-		kvm_flush_remote_tlbs(vcpu->kvm);
-
-	return nr_present;
+	return set_spte_ret & SET_SPTE_NEED_REMOTE_TLB_FLUSH;
 }
 
 #undef pt_element_t