KVM: x86: enable dirty log gradually in small chunks

It could take kvm->mmu_lock for an extended period of time when
enabling dirty log for the first time. The main cost is to clear
all the D-bits of last level SPTEs. This situation can benefit from
manual dirty log protect as well, which can reduce the mmu_lock
time taken. The sequence is like this:

1. Initialize all the bits of the dirty bitmap to 1 when enabling
   dirty log for the first time
2. Only write protect the huge pages
3. KVM_GET_DIRTY_LOG returns the dirty bitmap info
4. KVM_CLEAR_DIRTY_LOG will clear D-bit for each of the leaf level
   SPTEs gradually in small chunks

Under the Intel(R) Xeon(R) Gold 6152 CPU @ 2.10GHz environment,
I did some tests with a 128G windows VM and counted the time taken
of memory_global_dirty_log_start, here is the numbers:

VM Size        Before    After optimization
128G           460ms     10ms

Signed-off-by: Jay Zhou <jianjay.zhou@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 17 insertions, 7 deletions

diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 26ccb6c0a461..699ff9b35c88 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -858,7 +858,7 @@ static int kvm_vm_release(struct inode *inode, struct file *filp)
  * Allocation size is twice as large as the actual dirty bitmap size.
  * See kvm_vm_ioctl_get_dirty_log() why this is needed.
  */
-static int kvm_create_dirty_bitmap(struct kvm_memory_slot *memslot)
+static int kvm_alloc_dirty_bitmap(struct kvm_memory_slot *memslot)
 {
 	unsigned long dirty_bytes = 2 * kvm_dirty_bitmap_bytes(memslot);
 
@@ -1288,9 +1288,12 @@ int __kvm_set_memory_region(struct kvm *kvm,
 	if (!(new.flags & KVM_MEM_LOG_DIRTY_PAGES))
 		new.dirty_bitmap = NULL;
 	else if (!new.dirty_bitmap) {
-		r = kvm_create_dirty_bitmap(&new);
+		r = kvm_alloc_dirty_bitmap(&new);
 		if (r)
 			return r;
+
+		if (kvm_dirty_log_manual_protect_and_init_set(kvm))
+			bitmap_set(new.dirty_bitmap, 0, new.npages);
 	}
 
 	r = kvm_set_memslot(kvm, mem, &old, &new, as_id, change);
@@ -3529,9 +3532,6 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
 	case KVM_CAP_IOEVENTFD_ANY_LENGTH:
 	case KVM_CAP_CHECK_EXTENSION_VM:
 	case KVM_CAP_ENABLE_CAP_VM:
-#ifdef CONFIG_KVM_GENERIC_DIRTYLOG_READ_PROTECT
-	case KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2:
-#endif
 		return 1;
 #ifdef CONFIG_KVM_MMIO
 	case KVM_CAP_COALESCED_MMIO:
@@ -3539,6 +3539,10 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
 	case KVM_CAP_COALESCED_PIO:
 		return 1;
 #endif
+#ifdef CONFIG_KVM_GENERIC_DIRTYLOG_READ_PROTECT
+	case KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2:
+		return KVM_DIRTY_LOG_MANUAL_CAPS;
+#endif
 #ifdef CONFIG_HAVE_KVM_IRQ_ROUTING
 	case KVM_CAP_IRQ_ROUTING:
 		return KVM_MAX_IRQ_ROUTES;
@@ -3566,11 +3570,17 @@ static int kvm_vm_ioctl_enable_cap_generic(struct kvm *kvm,
 {
 	switch (cap->cap) {
 #ifdef CONFIG_KVM_GENERIC_DIRTYLOG_READ_PROTECT
-	case KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2:
-		if (cap->flags || (cap->args[0] & ~1))
+	case KVM_CAP_MANUAL_DIRTY_LOG_PROTECT2: {
+		u64 allowed_options = KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE;
+
+		if (cap->args[0] & KVM_DIRTY_LOG_MANUAL_PROTECT_ENABLE)
+			allowed_options = KVM_DIRTY_LOG_MANUAL_CAPS;
+
+		if (cap->flags || (cap->args[0] & ~allowed_options))
 			return -EINVAL;
 		kvm->manual_dirty_log_protect = cap->args[0];
 		return 0;
+	}
 #endif
 	default:
 		return kvm_vm_ioctl_enable_cap(kvm, cap);