1 files changed, 103 insertions, 22 deletions

diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 4ceb28675fdf..ad3a2bf2f1e9 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -30,6 +30,44 @@ static int so_priority;
 module_param(so_priority, int, 0644);
 MODULE_PARM_DESC(so_priority, "nvme tcp socket optimize priority");
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+/* lockdep can detect a circular dependency of the form
+ *   sk_lock -> mmap_lock (page fault) -> fs locks -> sk_lock
+ * because dependencies are tracked for both nvme-tcp and user contexts. Using
+ * a separate class prevents lockdep from conflating nvme-tcp socket use with
+ * user-space socket API use.
+ */
+static struct lock_class_key nvme_tcp_sk_key[2];
+static struct lock_class_key nvme_tcp_slock_key[2];
+
+static void nvme_tcp_reclassify_socket(struct socket *sock)
+{
+	struct sock *sk = sock->sk;
+
+	if (WARN_ON_ONCE(!sock_allow_reclassification(sk)))
+		return;
+
+	switch (sk->sk_family) {
+	case AF_INET:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET-NVME",
+					      &nvme_tcp_slock_key[0],
+					      "sk_lock-AF_INET-NVME",
+					      &nvme_tcp_sk_key[0]);
+		break;
+	case AF_INET6:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET6-NVME",
+					      &nvme_tcp_slock_key[1],
+					      "sk_lock-AF_INET6-NVME",
+					      &nvme_tcp_sk_key[1]);
+		break;
+	default:
+		WARN_ON_ONCE(1);
+	}
+}
+#else
+static void nvme_tcp_reclassify_socket(struct socket *sock) { }
+#endif
+
 enum nvme_tcp_send_state {
 	NVME_TCP_SEND_CMD_PDU = 0,
 	NVME_TCP_SEND_H2C_PDU,
@@ -44,6 +82,8 @@ struct nvme_tcp_request {
 	u32			data_len;
 	u32			pdu_len;
 	u32			pdu_sent;
+	u32			h2cdata_left;
+	u32			h2cdata_offset;
 	u16			ttag;
 	__le16			status;
 	struct list_head	entry;
@@ -95,6 +135,7 @@ struct nvme_tcp_queue {
 	struct nvme_tcp_request *request;
 
 	int			queue_size;
+	u32			maxh2cdata;
 	size_t			cmnd_capsule_len;
 	struct nvme_tcp_ctrl	*ctrl;
 	unsigned long		flags;
@@ -572,23 +613,26 @@ static int nvme_tcp_handle_comp(struct nvme_tcp_queue *queue,
 	return ret;
 }
 
-static void nvme_tcp_setup_h2c_data_pdu(struct nvme_tcp_request *req,
-		struct nvme_tcp_r2t_pdu *pdu)
+static void nvme_tcp_setup_h2c_data_pdu(struct nvme_tcp_request *req)
 {
 	struct nvme_tcp_data_pdu *data = req->pdu;
 	struct nvme_tcp_queue *queue = req->queue;
 	struct request *rq = blk_mq_rq_from_pdu(req);
+	u32 h2cdata_sent = req->pdu_len;
 	u8 hdgst = nvme_tcp_hdgst_len(queue);
 	u8 ddgst = nvme_tcp_ddgst_len(queue);
 
 	req->state = NVME_TCP_SEND_H2C_PDU;
 	req->offset = 0;
-	req->pdu_len = le32_to_cpu(pdu->r2t_length);
+	req->pdu_len = min(req->h2cdata_left, queue->maxh2cdata);
 	req->pdu_sent = 0;
+	req->h2cdata_left -= req->pdu_len;
+	req->h2cdata_offset += h2cdata_sent;
 
 	memset(data, 0, sizeof(*data));
 	data->hdr.type = nvme_tcp_h2c_data;
-	data->hdr.flags = NVME_TCP_F_DATA_LAST;
+	if (!req->h2cdata_left)
+		data->hdr.flags = NVME_TCP_F_DATA_LAST;
 	if (queue->hdr_digest)
 		data->hdr.flags |= NVME_TCP_F_HDGST;
 	if (queue->data_digest)
@@ -597,9 +641,9 @@ static void nvme_tcp_setup_h2c_data_pdu(struct nvme_tcp_request *req,
 	data->hdr.pdo = data->hdr.hlen + hdgst;
 	data->hdr.plen =
 		cpu_to_le32(data->hdr.hlen + hdgst + req->pdu_len + ddgst);
-	data->ttag = pdu->ttag;
+	data->ttag = req->ttag;
 	data->command_id = nvme_cid(rq);
-	data->data_offset = pdu->r2t_offset;
+	data->data_offset = cpu_to_le32(req->h2cdata_offset);
 	data->data_length = cpu_to_le32(req->pdu_len);
 }
 
@@ -609,6 +653,7 @@ static int nvme_tcp_handle_r2t(struct nvme_tcp_queue *queue,
 	struct nvme_tcp_request *req;
 	struct request *rq;
 	u32 r2t_length = le32_to_cpu(pdu->r2t_length);
+	u32 r2t_offset = le32_to_cpu(pdu->r2t_offset);
 
 	rq = nvme_find_rq(nvme_tcp_tagset(queue), pdu->command_id);
 	if (!rq) {
@@ -633,14 +678,19 @@ static int nvme_tcp_handle_r2t(struct nvme_tcp_queue *queue,
 		return -EPROTO;
 	}
 
-	if (unlikely(le32_to_cpu(pdu->r2t_offset) < req->data_sent)) {
+	if (unlikely(r2t_offset < req->data_sent)) {
 		dev_err(queue->ctrl->ctrl.device,
 			"req %d unexpected r2t offset %u (expected %zu)\n",
-			rq->tag, le32_to_cpu(pdu->r2t_offset), req->data_sent);
+			rq->tag, r2t_offset, req->data_sent);
 		return -EPROTO;
 	}
 
-	nvme_tcp_setup_h2c_data_pdu(req, pdu);
+	req->pdu_len = 0;
+	req->h2cdata_left = r2t_length;
+	req->h2cdata_offset = r2t_offset;
+	req->ttag = pdu->ttag;
+
+	nvme_tcp_setup_h2c_data_pdu(req);
 	nvme_tcp_queue_request(req, false, true);
 
 	return 0;
@@ -913,13 +963,22 @@ static inline void nvme_tcp_done_send_req(struct nvme_tcp_queue *queue)
 
 static void nvme_tcp_fail_request(struct nvme_tcp_request *req)
 {
-	nvme_tcp_end_request(blk_mq_rq_from_pdu(req), NVME_SC_HOST_PATH_ERROR);
+	if (nvme_tcp_async_req(req)) {
+		union nvme_result res = {};
+
+		nvme_complete_async_event(&req->queue->ctrl->ctrl,
+				cpu_to_le16(NVME_SC_HOST_PATH_ERROR), &res);
+	} else {
+		nvme_tcp_end_request(blk_mq_rq_from_pdu(req),
+				NVME_SC_HOST_PATH_ERROR);
+	}
 }
 
 static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
 {
 	struct nvme_tcp_queue *queue = req->queue;
 	int req_data_len = req->data_len;
+	u32 h2cdata_left = req->h2cdata_left;
 
 	while (true) {
 		struct page *page = nvme_tcp_req_cur_page(req);
@@ -964,7 +1023,10 @@ static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
 				req->state = NVME_TCP_SEND_DDGST;
 				req->offset = 0;
 			} else {
-				nvme_tcp_done_send_req(queue);
+				if (h2cdata_left)
+					nvme_tcp_setup_h2c_data_pdu(req);
+				else
+					nvme_tcp_done_send_req(queue);
 			}
 			return 1;
 		}
@@ -1022,9 +1084,14 @@ static int nvme_tcp_try_send_data_pdu(struct nvme_tcp_request *req)
 	if (queue->hdr_digest && !req->offset)
 		nvme_tcp_hdgst(queue->snd_hash, pdu, sizeof(*pdu));
 
-	ret = kernel_sendpage(queue->sock, virt_to_page(pdu),
-			offset_in_page(pdu) + req->offset, len,
-			MSG_DONTWAIT | MSG_MORE | MSG_SENDPAGE_NOTLAST);
+	if (!req->h2cdata_left)
+		ret = kernel_sendpage(queue->sock, virt_to_page(pdu),
+				offset_in_page(pdu) + req->offset, len,
+				MSG_DONTWAIT | MSG_MORE | MSG_SENDPAGE_NOTLAST);
+	else
+		ret = sock_no_sendpage(queue->sock, virt_to_page(pdu),
+				offset_in_page(pdu) + req->offset, len,
+				MSG_DONTWAIT | MSG_MORE);
 	if (unlikely(ret <= 0))
 		return ret;
 
@@ -1044,6 +1111,7 @@ static int nvme_tcp_try_send_ddgst(struct nvme_tcp_request *req)
 {
 	struct nvme_tcp_queue *queue = req->queue;
 	size_t offset = req->offset;
+	u32 h2cdata_left = req->h2cdata_left;
 	int ret;
 	struct msghdr msg = { .msg_flags = MSG_DONTWAIT };
 	struct kvec iov = {
@@ -1061,7 +1129,10 @@ static int nvme_tcp_try_send_ddgst(struct nvme_tcp_request *req)
 		return ret;
 
 	if (offset + ret == NVME_TCP_DIGEST_LENGTH) {
-		nvme_tcp_done_send_req(queue);
+		if (h2cdata_left)
+			nvme_tcp_setup_h2c_data_pdu(req);
+		else
+			nvme_tcp_done_send_req(queue);
 		return 1;
 	}
 
@@ -1253,6 +1324,7 @@ static int nvme_tcp_init_connection(struct nvme_tcp_queue *queue)
 	struct msghdr msg = {};
 	struct kvec iov;
 	bool ctrl_hdgst, ctrl_ddgst;
+	u32 maxh2cdata;
 	int ret;
 
 	icreq = kzalloc(sizeof(*icreq), GFP_KERNEL);
@@ -1336,6 +1408,14 @@ static int nvme_tcp_init_connection(struct nvme_tcp_queue *queue)
 		goto free_icresp;
 	}
 
+	maxh2cdata = le32_to_cpu(icresp->maxdata);
+	if ((maxh2cdata % 4) || (maxh2cdata < NVME_TCP_MIN_MAXH2CDATA)) {
+		pr_err("queue %d: invalid maxh2cdata returned %u\n",
+		       nvme_tcp_queue_id(queue), maxh2cdata);
+		goto free_icresp;
+	}
+	queue->maxh2cdata = maxh2cdata;
+
 	ret = 0;
 free_icresp:
 	kfree(icresp);
@@ -1427,6 +1507,8 @@ static int nvme_tcp_alloc_queue(struct nvme_ctrl *nctrl,
 		goto err_destroy_mutex;
 	}
 
+	nvme_tcp_reclassify_socket(queue->sock);
+
 	/* Single syn retry */
 	tcp_sock_set_syncnt(queue->sock->sk, 1);
 
@@ -1674,7 +1756,7 @@ static void nvme_tcp_stop_io_queues(struct nvme_ctrl *ctrl)
 
 static int nvme_tcp_start_io_queues(struct nvme_ctrl *ctrl)
 {
-	int i, ret = 0;
+	int i, ret;
 
 	for (i = 1; i < ctrl->queue_count; i++) {
 		ret = nvme_tcp_start_queue(ctrl, i);
@@ -1714,8 +1796,7 @@ static int __nvme_tcp_alloc_io_queues(struct nvme_ctrl *ctrl)
 	int i, ret;
 
 	for (i = 1; i < ctrl->queue_count; i++) {
-		ret = nvme_tcp_alloc_queue(ctrl, i,
-				ctrl->sqsize + 1);
+		ret = nvme_tcp_alloc_queue(ctrl, i, ctrl->sqsize + 1);
 		if (ret)
 			goto out_free_queues;
 	}
@@ -1825,11 +1906,9 @@ static int nvme_tcp_configure_io_queues(struct nvme_ctrl *ctrl, bool new)
 			goto out_free_io_queues;
 		}
 
-		ctrl->connect_q = blk_mq_init_queue(ctrl->tagset);
-		if (IS_ERR(ctrl->connect_q)) {
-			ret = PTR_ERR(ctrl->connect_q);
+		ret = nvme_ctrl_init_connect_q(ctrl);
+		if (ret)
 			goto out_free_tag_set;
-		}
 	}
 
 	ret = nvme_tcp_start_io_queues(ctrl);
@@ -2096,6 +2175,7 @@ static void nvme_tcp_error_recovery_work(struct work_struct *work)
 	struct nvme_ctrl *ctrl = &tcp_ctrl->ctrl;
 
 	nvme_stop_keep_alive(ctrl);
+	flush_work(&ctrl->async_event_work);
 	nvme_tcp_teardown_io_queues(ctrl, false);
 	/* unquiesce to fail fast pending requests */
 	nvme_start_queues(ctrl);
@@ -2320,6 +2400,7 @@ static blk_status_t nvme_tcp_setup_cmd_pdu(struct nvme_ns *ns,
 	req->data_sent = 0;
 	req->pdu_len = 0;
 	req->pdu_sent = 0;
+	req->h2cdata_left = 0;
 	req->data_len = blk_rq_nr_phys_segments(rq) ?
 				blk_rq_payload_bytes(rq) : 0;
 	req->curr_bio = rq->bio;