vhost-vdpa: support IOTLB batching hints

This patches extend the vhost IOTLB API to accept batch updating hints
form userspace. When userspace wants update the device IOTLB in a
batch, it may do:

1) Write vhost_iotlb_msg with VHOST_IOTLB_BATCH_BEGIN flag
2) Perform a batch of IOTLB updating via VHOST_IOTLB_UPDATE/INVALIDATE
3) Write vhost_iotlb_msg with VHOST_IOTLB_BATCH_END flag

Vhost-vdpa may decide to batch the IOMMU/IOTLB updating in step 3 when
vDPA device support set_map() ops. This is useful for the vDPA device
that want to know all the mappings to tweak their own DMA translation
logic.

For vDPA device that doesn't require set_map(), no behavior changes.

This capability is advertised via VHOST_BACKEND_F_IOTLB_BATCH capability.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20200804162048.22587-5-eli@mellanox.comSigned-off-by: Michael S. Tsirkin <mst@redhat.com>

drivers/vhost/vdpa.c

@@ -27,7 +27,9 @@
 #include "vhost.h"
 
 enum {
-	VHOST_VDPA_BACKEND_FEATURES = (1ULL << VHOST_BACKEND_F_IOTLB_MSG_V2)
+	VHOST_VDPA_BACKEND_FEATURES =
+	(1ULL << VHOST_BACKEND_F_IOTLB_MSG_V2) |
+	(1ULL << VHOST_BACKEND_F_IOTLB_BATCH),
 };
 
 /* Currently, only network backend w/o multiqueue is supported. */
@@ -48,6 +50,7 @@ struct vhost_vdpa {
 	int virtio_id;
 	int minor;
 	struct eventfd_ctx *config_ctx;
+	int in_batch;
 };
 
 static DEFINE_IDA(vhost_vdpa_ida);
@@ -124,6 +127,7 @@ static void vhost_vdpa_reset(struct vhost_vdpa *v)
 	struct vdpa_device *vdpa = v->vdpa;
 
 	vdpa_reset(vdpa);
+	v->in_batch = 0;
 }
 
 static long vhost_vdpa_get_device_id(struct vhost_vdpa *v, u8 __user *argp)
@@ -546,13 +550,15 @@ static int vhost_vdpa_map(struct vhost_vdpa *v,
 	if (r)
 		return r;
 
-	if (ops->dma_map)
+	if (ops->dma_map) {
 		r = ops->dma_map(vdpa, iova, size, pa, perm);
-	else if (ops->set_map)
+	} else if (ops->set_map) {
+		if (!v->in_batch)
 			r = ops->set_map(vdpa, dev->iotlb);
-	else
+	} else {
 		r = iommu_map(v->domain, iova, pa, size,
 			      perm_to_iommu_flags(perm));
+	}
 
 	return r;
 }
@@ -565,12 +571,14 @@ static void vhost_vdpa_unmap(struct vhost_vdpa *v, u64 iova, u64 size)
 
 	vhost_vdpa_iotlb_unmap(v, iova, iova + size - 1);
 
-	if (ops->dma_map)
+	if (ops->dma_map) {
 		ops->dma_unmap(vdpa, iova, size);
-	else if (ops->set_map)
+	} else if (ops->set_map) {
+		if (!v->in_batch)
 			ops->set_map(vdpa, dev->iotlb);
-	else
+	} else {
 		iommu_unmap(v->domain, iova, size);
+	}
 }
 
 static int vhost_vdpa_process_iotlb_update(struct vhost_vdpa *v,
@@ -663,6 +671,8 @@ static int vhost_vdpa_process_iotlb_msg(struct vhost_dev *dev,
 					struct vhost_iotlb_msg *msg)
 {
 	struct vhost_vdpa *v = container_of(dev, struct vhost_vdpa, vdev);
+	struct vdpa_device *vdpa = v->vdpa;
+	const struct vdpa_config_ops *ops = vdpa->config;
 	int r = 0;
 
 	r = vhost_dev_check_owner(dev);
@@ -676,6 +686,14 @@ static int vhost_vdpa_process_iotlb_msg(struct vhost_dev *dev,
 	case VHOST_IOTLB_INVALIDATE:
 		vhost_vdpa_unmap(v, msg->iova, msg->size);
 		break;
+	case VHOST_IOTLB_BATCH_BEGIN:
+		v->in_batch = true;
+		break;
+	case VHOST_IOTLB_BATCH_END:
+		if (v->in_batch && ops->set_map)
+			ops->set_map(vdpa, dev->iotlb);
+		v->in_batch = false;
+		break;
 	default:
 		r = -EINVAL;
 		break;

include/uapi/linux/vhost.h

@@ -91,6 +91,8 @@
 
 /* Use message type V2 */
 #define VHOST_BACKEND_F_IOTLB_MSG_V2 0x1
+/* IOTLB can accept batching hints */
+#define VHOST_BACKEND_F_IOTLB_BATCH  0x2
 
 #define VHOST_SET_BACKEND_FEATURES _IOW(VHOST_VIRTIO, 0x25, __u64)
 #define VHOST_GET_BACKEND_FEATURES _IOR(VHOST_VIRTIO, 0x26, __u64)

include/uapi/linux/vhost_types.h

@@ -60,6 +60,17 @@ struct vhost_iotlb_msg {
 #define VHOST_IOTLB_UPDATE         2
 #define VHOST_IOTLB_INVALIDATE     3
 #define VHOST_IOTLB_ACCESS_FAIL    4
+/*
+ * VHOST_IOTLB_BATCH_BEGIN and VHOST_IOTLB_BATCH_END allow modifying
+ * multiple mappings in one go: beginning with
+ * VHOST_IOTLB_BATCH_BEGIN, followed by any number of
+ * VHOST_IOTLB_UPDATE messages, and ending with VHOST_IOTLB_BATCH_END.
+ * When one of these two values is used as the message type, the rest
+ * of the fields in the message are ignored. There's no guarantee that
+ * these changes take place automatically in the device.
+ */
+#define VHOST_IOTLB_BATCH_BEGIN    5
+#define VHOST_IOTLB_BATCH_END      6
 	__u8 type;
 };