KVM: MMU: delay flush all tlbs on sync_page path

Quote from Avi:
| I don't think we need to flush immediately; set a "tlb dirty" bit somewhere
| that is cleareded when we flush the tlb.  kvm_mmu_notifier_invalidate_page()
| can consult the bit and force a flush if set.

Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 5156d45..ee99b77 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -168,8 +168,12 @@
 
 void kvm_flush_remote_tlbs(struct kvm *kvm)
 {
+	int dirty_count = kvm->tlbs_dirty;
+
+	smp_mb();
 	if (make_all_cpus_request(kvm, KVM_REQ_TLB_FLUSH))
 		++kvm->stat.remote_tlb_flush;
+	cmpxchg(&kvm->tlbs_dirty, dirty_count, 0);
 }
 
 void kvm_reload_remote_mmus(struct kvm *kvm)
@@ -249,7 +253,7 @@
 	idx = srcu_read_lock(&kvm->srcu);
 	spin_lock(&kvm->mmu_lock);
 	kvm->mmu_notifier_seq++;
-	need_tlb_flush = kvm_unmap_hva(kvm, address);
+	need_tlb_flush = kvm_unmap_hva(kvm, address) | kvm->tlbs_dirty;
 	spin_unlock(&kvm->mmu_lock);
 	srcu_read_unlock(&kvm->srcu, idx);
 
@@ -293,6 +297,7 @@
 	kvm->mmu_notifier_count++;
 	for (; start < end; start += PAGE_SIZE)
 		need_tlb_flush |= kvm_unmap_hva(kvm, start);
+	need_tlb_flush |= kvm->tlbs_dirty;
 	spin_unlock(&kvm->mmu_lock);
 	srcu_read_unlock(&kvm->srcu, idx);