1 files changed, 97 insertions, 20 deletions

diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 6d43b23a0fc8..0ef14f0fad86 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -20,6 +20,16 @@
 
 struct nvme_tcp_queue;
 
+/* Define the socket priority to use for connections were it is desirable
+ * that the NIC consider performing optimized packet processing or filtering.
+ * A non-zero value being sufficient to indicate general consideration of any
+ * possible optimization.  Making it a module param allows for alternative
+ * values that may be unique for some NIC implementations.
+ */
+static int so_priority;
+module_param(so_priority, int, 0644);
+MODULE_PARM_DESC(so_priority, "nvme tcp socket optimize priority");
+
 enum nvme_tcp_send_state {
 	NVME_TCP_SEND_CMD_PDU = 0,
 	NVME_TCP_SEND_H2C_PDU,
@@ -422,7 +432,7 @@ static void nvme_tcp_error_recovery(struct nvme_ctrl *ctrl)
 	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_RESETTING))
 		return;
 
-	queue_work(nvme_wq, &to_tcp_ctrl(ctrl)->err_work);
+	queue_work(nvme_reset_wq, &to_tcp_ctrl(ctrl)->err_work);
 }
 
 static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
@@ -1017,8 +1027,15 @@ static int nvme_tcp_try_send(struct nvme_tcp_queue *queue)
 	if (req->state == NVME_TCP_SEND_DDGST)
 		ret = nvme_tcp_try_send_ddgst(req);
 done:
-	if (ret == -EAGAIN)
+	if (ret == -EAGAIN) {
 		ret = 0;
+	} else if (ret < 0) {
+		dev_err(queue->ctrl->ctrl.device,
+			"failed to send request %d\n", ret);
+		if (ret != -EPIPE && ret != -ECONNRESET)
+			nvme_tcp_fail_request(queue->request);
+		nvme_tcp_done_send_req(queue);
+	}
 	return ret;
 }
 
@@ -1049,20 +1066,16 @@ static void nvme_tcp_io_work(struct work_struct *w)
 		int result;
 
 		result = nvme_tcp_try_send(queue);
-		if (result > 0) {
+		if (result > 0)
 			pending = true;
-		} else if (unlikely(result < 0)) {
-			dev_err(queue->ctrl->ctrl.device,
-				"failed to send request %d\n", result);
-			if (result != -EPIPE)
-				nvme_tcp_fail_request(queue->request);
-			nvme_tcp_done_send_req(queue);
-			return;
-		}
+		else if (unlikely(result < 0))
+			break;
 
 		result = nvme_tcp_try_recv(queue);
 		if (result > 0)
 			pending = true;
+		else if (unlikely(result < 0))
+			break;
 
 		if (!pending)
 			return;
@@ -1243,13 +1256,67 @@ free_icreq:
 	return ret;
 }
 
+static bool nvme_tcp_admin_queue(struct nvme_tcp_queue *queue)
+{
+	return nvme_tcp_queue_id(queue) == 0;
+}
+
+static bool nvme_tcp_default_queue(struct nvme_tcp_queue *queue)
+{
+	struct nvme_tcp_ctrl *ctrl = queue->ctrl;
+	int qid = nvme_tcp_queue_id(queue);
+
+	return !nvme_tcp_admin_queue(queue) &&
+		qid < 1 + ctrl->io_queues[HCTX_TYPE_DEFAULT];
+}
+
+static bool nvme_tcp_read_queue(struct nvme_tcp_queue *queue)
+{
+	struct nvme_tcp_ctrl *ctrl = queue->ctrl;
+	int qid = nvme_tcp_queue_id(queue);
+
+	return !nvme_tcp_admin_queue(queue) &&
+		!nvme_tcp_default_queue(queue) &&
+		qid < 1 + ctrl->io_queues[HCTX_TYPE_DEFAULT] +
+			  ctrl->io_queues[HCTX_TYPE_READ];
+}
+
+static bool nvme_tcp_poll_queue(struct nvme_tcp_queue *queue)
+{
+	struct nvme_tcp_ctrl *ctrl = queue->ctrl;
+	int qid = nvme_tcp_queue_id(queue);
+
+	return !nvme_tcp_admin_queue(queue) &&
+		!nvme_tcp_default_queue(queue) &&
+		!nvme_tcp_read_queue(queue) &&
+		qid < 1 + ctrl->io_queues[HCTX_TYPE_DEFAULT] +
+			  ctrl->io_queues[HCTX_TYPE_READ] +
+			  ctrl->io_queues[HCTX_TYPE_POLL];
+}
+
+static void nvme_tcp_set_queue_io_cpu(struct nvme_tcp_queue *queue)
+{
+	struct nvme_tcp_ctrl *ctrl = queue->ctrl;
+	int qid = nvme_tcp_queue_id(queue);
+	int n = 0;
+
+	if (nvme_tcp_default_queue(queue))
+		n = qid - 1;
+	else if (nvme_tcp_read_queue(queue))
+		n = qid - ctrl->io_queues[HCTX_TYPE_DEFAULT] - 1;
+	else if (nvme_tcp_poll_queue(queue))
+		n = qid - ctrl->io_queues[HCTX_TYPE_DEFAULT] -
+				ctrl->io_queues[HCTX_TYPE_READ] - 1;
+	queue->io_cpu = cpumask_next_wrap(n - 1, cpu_online_mask, -1, false);
+}
+
 static int nvme_tcp_alloc_queue(struct nvme_ctrl *nctrl,
 		int qid, size_t queue_size)
 {
 	struct nvme_tcp_ctrl *ctrl = to_tcp_ctrl(nctrl);
 	struct nvme_tcp_queue *queue = &ctrl->queues[qid];
 	struct linger sol = { .l_onoff = 1, .l_linger = 0 };
-	int ret, opt, rcv_pdu_size, n;
+	int ret, opt, rcv_pdu_size;
 
 	queue->ctrl = ctrl;
 	INIT_LIST_HEAD(&queue->send_list);
@@ -1304,6 +1371,17 @@ static int nvme_tcp_alloc_queue(struct nvme_ctrl *nctrl,
 		goto err_sock;
 	}
 
+	if (so_priority > 0) {
+		ret = kernel_setsockopt(queue->sock, SOL_SOCKET, SO_PRIORITY,
+				(char *)&so_priority, sizeof(so_priority));
+		if (ret) {
+			dev_err(ctrl->ctrl.device,
+				"failed to set SO_PRIORITY sock opt, ret %d\n",
+				ret);
+			goto err_sock;
+		}
+	}
+
 	/* Set socket type of service */
 	if (nctrl->opts->tos >= 0) {
 		opt = nctrl->opts->tos;
@@ -1317,11 +1395,7 @@ static int nvme_tcp_alloc_queue(struct nvme_ctrl *nctrl,
 	}
 
 	queue->sock->sk->sk_allocation = GFP_ATOMIC;
-	if (!qid)
-		n = 0;
-	else
-		n = (qid - 1) % num_online_cpus();
-	queue->io_cpu = cpumask_next_wrap(n - 1, cpu_online_mask, -1, false);
+	nvme_tcp_set_queue_io_cpu(queue);
 	queue->request = NULL;
 	queue->data_remaining = 0;
 	queue->ddgst_remaining = 0;
@@ -1856,8 +1930,13 @@ static int nvme_tcp_setup_ctrl(struct nvme_ctrl *ctrl, bool new)
 	}
 
 	if (!nvme_change_ctrl_state(ctrl, NVME_CTRL_LIVE)) {
-		/* state change failure is ok if we're in DELETING state */
+		/*
+		 * state change failure is ok if we're in DELETING state,
+		 * unless we're during creation of a new controller to
+		 * avoid races with teardown flow.
+		 */
 		WARN_ON_ONCE(ctrl->state != NVME_CTRL_DELETING);
+		WARN_ON_ONCE(new);
 		ret = -EINVAL;
 		goto destroy_io;
 	}
@@ -2354,8 +2433,6 @@ static struct nvme_ctrl *nvme_tcp_create_ctrl(struct device *dev,
 	dev_info(ctrl->ctrl.device, "new ctrl: NQN \"%s\", addr %pISp\n",
 		ctrl->ctrl.opts->subsysnqn, &ctrl->addr);
 
-	nvme_get_ctrl(&ctrl->ctrl);
-
 	mutex_lock(&nvme_tcp_ctrl_mutex);
 	list_add_tail(&ctrl->list, &nvme_tcp_ctrl_list);
 	mutex_unlock(&nvme_tcp_ctrl_mutex);