Merge branch 'nvme-4.16' of git://git.infradead.org/nvme into for-4.16/block

Pull NVMe fixes from Christoph:

"Below are the pending nvme updates for Linux 4.16. Just fixes and
 cleanups from various contributors this time around."

drivers/nvme/host/core.c

@@ -232,6 +232,15 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 
 	old_state = ctrl->state;
 	switch (new_state) {
+	case NVME_CTRL_ADMIN_ONLY:
+		switch (old_state) {
+		case NVME_CTRL_RESETTING:
+			changed = true;
+			/* FALLTHRU */
+		default:
+			break;
+		}
+		break;
 	case NVME_CTRL_LIVE:
 		switch (old_state) {
 		case NVME_CTRL_NEW:
@@ -247,6 +256,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 		switch (old_state) {
 		case NVME_CTRL_NEW:
 		case NVME_CTRL_LIVE:
+		case NVME_CTRL_ADMIN_ONLY:
 			changed = true;
 			/* FALLTHRU */
 		default:
@@ -266,6 +276,7 @@ bool nvme_change_ctrl_state(struct nvme_ctrl *ctrl,
 	case NVME_CTRL_DELETING:
 		switch (old_state) {
 		case NVME_CTRL_LIVE:
+		case NVME_CTRL_ADMIN_ONLY:
 		case NVME_CTRL_RESETTING:
 		case NVME_CTRL_RECONNECTING:
 			changed = true;
@@ -1217,16 +1228,27 @@ static int nvme_open(struct block_device *bdev, fmode_t mode)
 #ifdef CONFIG_NVME_MULTIPATH
 	/* should never be called due to GENHD_FL_HIDDEN */
 	if (WARN_ON_ONCE(ns->head->disk))
-		return -ENXIO;
+		goto fail;
 #endif
 	if (!kref_get_unless_zero(&ns->kref))
-		return -ENXIO;
+		goto fail;
+	if (!try_module_get(ns->ctrl->ops->module))
+		goto fail_put_ns;
+
 	return 0;
+
+fail_put_ns:
+	nvme_put_ns(ns);
+fail:
+	return -ENXIO;
 }
 
 static void nvme_release(struct gendisk *disk, fmode_t mode)
 {
-	nvme_put_ns(disk->private_data);
+	struct nvme_ns *ns = disk->private_data;
+
+	module_put(ns->ctrl->ops->module);
+	nvme_put_ns(ns);
 }
 
 static int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo)
@@ -2047,6 +2069,22 @@ static const struct attribute_group *nvme_subsys_attrs_groups[] = {
 	NULL,
 };
 
+static int nvme_active_ctrls(struct nvme_subsystem *subsys)
+{
+	int count = 0;
+	struct nvme_ctrl *ctrl;
+
+	mutex_lock(&subsys->lock);
+	list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) {
+		if (ctrl->state != NVME_CTRL_DELETING &&
+		    ctrl->state != NVME_CTRL_DEAD)
+			count++;
+	}
+	mutex_unlock(&subsys->lock);
+
+	return count;
+}
+
 static int nvme_init_subsystem(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 {
 	struct nvme_subsystem *subsys, *found;
@@ -2085,7 +2123,7 @@ static int nvme_init_subsystem(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 		 * Verify that the subsystem actually supports multiple
 		 * controllers, else bail out.
 		 */
-		if (!(id->cmic & (1 << 1))) {
+		if (nvme_active_ctrls(found) && !(id->cmic & (1 << 1))) {
 			dev_err(ctrl->device,
 				"ignoring ctrl due to duplicate subnqn (%s).\n",
 				found->subnqn);
@@ -2252,7 +2290,7 @@ int nvme_init_identify(struct nvme_ctrl *ctrl)
 						 shutdown_timeout, 60);
 
 		if (ctrl->shutdown_timeout != shutdown_timeout)
-			dev_warn(ctrl->device,
+			dev_info(ctrl->device,
 				 "Shutdown timeout set to %u seconds\n",
 				 ctrl->shutdown_timeout);
 	} else
@@ -2336,8 +2374,14 @@ static int nvme_dev_open(struct inode *inode, struct file *file)
 	struct nvme_ctrl *ctrl =
 		container_of(inode->i_cdev, struct nvme_ctrl, cdev);
 
-	if (ctrl->state != NVME_CTRL_LIVE)
+	switch (ctrl->state) {
+	case NVME_CTRL_LIVE:
+	case NVME_CTRL_ADMIN_ONLY:
+		break;
+	default:
 		return -EWOULDBLOCK;
+	}
+
 	file->private_data = ctrl;
 	return 0;
 }
@@ -2601,6 +2645,7 @@ static ssize_t nvme_sysfs_show_state(struct device *dev,
 	static const char *const state_name[] = {
 		[NVME_CTRL_NEW]		= "new",
 		[NVME_CTRL_LIVE]	= "live",
+		[NVME_CTRL_ADMIN_ONLY]	= "only-admin",
 		[NVME_CTRL_RESETTING]	= "resetting",
 		[NVME_CTRL_RECONNECTING]= "reconnecting",
 		[NVME_CTRL_DELETING]	= "deleting",
@@ -3073,6 +3118,8 @@ static void nvme_scan_work(struct work_struct *work)
 	if (ctrl->state != NVME_CTRL_LIVE)
 		return;
 
+	WARN_ON_ONCE(!ctrl->tagset);
+
 	if (nvme_identify_ctrl(ctrl, &id))
 		return;
 
@@ -3093,8 +3140,7 @@ static void nvme_scan_work(struct work_struct *work)
 void nvme_queue_scan(struct nvme_ctrl *ctrl)
 {
 	/*
-	 * Do not queue new scan work when a controller is reset during
-	 * removal.
+	 * Only new queue scan work when admin and IO queues are both alive
 	 */
 	if (ctrl->state == NVME_CTRL_LIVE)
 		queue_work(nvme_wq, &ctrl->scan_work);

drivers/nvme/host/fabrics.c

@@ -492,7 +492,7 @@ EXPORT_SYMBOL_GPL(nvmf_should_reconnect);
  */
 int nvmf_register_transport(struct nvmf_transport_ops *ops)
 {
-	if (!ops->create_ctrl)
+	if (!ops->create_ctrl || !ops->module)
 		return -EINVAL;
 
 	down_write(&nvmf_transports_rwsem);
@@ -868,32 +868,41 @@ nvmf_create_ctrl(struct device *dev, const char *buf, size_t count)
 		goto out_unlock;
 	}
 
+	if (!try_module_get(ops->module)) {
+		ret = -EBUSY;
+		goto out_unlock;
+	}
+
 	ret = nvmf_check_required_opts(opts, ops->required_opts);
 	if (ret)
-		goto out_unlock;
+		goto out_module_put;
 	ret = nvmf_check_allowed_opts(opts, NVMF_ALLOWED_OPTS |
 				ops->allowed_opts | ops->required_opts);
 	if (ret)
-		goto out_unlock;
+		goto out_module_put;
 
 	ctrl = ops->create_ctrl(dev, opts);
 	if (IS_ERR(ctrl)) {
 		ret = PTR_ERR(ctrl);
-		goto out_unlock;
+		goto out_module_put;
 	}
 
 	if (strcmp(ctrl->subsys->subnqn, opts->subsysnqn)) {
 		dev_warn(ctrl->device,
 			"controller returned incorrect NQN: \"%s\".\n",
 			ctrl->subsys->subnqn);
+		module_put(ops->module);
 		up_read(&nvmf_transports_rwsem);
 		nvme_delete_ctrl_sync(ctrl);
 		return ERR_PTR(-EINVAL);
 	}
 
+	module_put(ops->module);
 	up_read(&nvmf_transports_rwsem);
 	return ctrl;
 
+out_module_put:
+	module_put(ops->module);
 out_unlock:
 	up_read(&nvmf_transports_rwsem);
 out_free_opts:

drivers/nvme/host/fabrics.h

@@ -108,6 +108,7 @@ struct nvmf_ctrl_options {
  *			       fabric implementation of NVMe fabrics.
  * @entry:		Used by the fabrics library to add the new
  *			registration entry to its linked-list internal tree.
+ * @module:             Transport module reference
  * @name:		Name of the NVMe fabric driver implementation.
  * @required_opts:	sysfs command-line options that must be specified
  *			when adding a new NVMe controller.
@@ -126,6 +127,7 @@ struct nvmf_ctrl_options {
  */
 struct nvmf_transport_ops {
 	struct list_head	entry;
+	struct module		*module;
 	const char		*name;
 	int			required_opts;
 	int			allowed_opts;

drivers/nvme/host/fc.c

@@ -3381,6 +3381,7 @@ nvme_fc_create_ctrl(struct device *dev, struct nvmf_ctrl_options *opts)
 
 static struct nvmf_transport_ops nvme_fc_transport = {
 	.name		= "fc",
+	.module		= THIS_MODULE,
 	.required_opts	= NVMF_OPT_TRADDR | NVMF_OPT_HOST_TRADDR,
 	.allowed_opts	= NVMF_OPT_RECONNECT_DELAY | NVMF_OPT_CTRL_LOSS_TMO,
 	.create_ctrl	= nvme_fc_create_ctrl,

drivers/nvme/host/nvme.h

@@ -119,6 +119,7 @@ static inline struct nvme_request *nvme_req(struct request *req)
 enum nvme_ctrl_state {
 	NVME_CTRL_NEW,
 	NVME_CTRL_LIVE,
+	NVME_CTRL_ADMIN_ONLY,    /* Only admin queue live */
 	NVME_CTRL_RESETTING,
 	NVME_CTRL_RECONNECTING,
 	NVME_CTRL_DELETING,

drivers/nvme/host/pci.c

@@ -1770,7 +1770,7 @@ static int __nvme_alloc_host_mem(struct nvme_dev *dev, u64 preferred,
 	dma_addr_t descs_dma;
 	int i = 0;
 	void **bufs;
-	u64 size = 0, tmp;
+	u64 size, tmp;
 
 	tmp = (preferred + chunk_size - 1);
 	do_div(tmp, chunk_size);
@@ -1853,7 +1853,7 @@ static int nvme_setup_host_mem(struct nvme_dev *dev)
 	u64 preferred = (u64)dev->ctrl.hmpre * 4096;
 	u64 min = (u64)dev->ctrl.hmmin * 4096;
 	u32 enable_bits = NVME_HOST_MEM_ENABLE;
-	int ret = 0;
+	int ret;
 
 	preferred = min(preferred, max);
 	if (min > max) {
@@ -2035,13 +2035,12 @@ static void nvme_disable_io_queues(struct nvme_dev *dev, int queues)
 }
 
 /*
- * Return: error value if an error occurred setting up the queues or calling
- * Identify Device.  0 if these succeeded, even if adding some of the
- * namespaces failed.  At the moment, these failures are silent.  TBD which
- * failures should be reported.
+ * return error value only when tagset allocation failed
  */
 static int nvme_dev_add(struct nvme_dev *dev)
 {
+	int ret;
+
 	if (!dev->ctrl.tagset) {
 		dev->tagset.ops = &nvme_mq_ops;
 		dev->tagset.nr_hw_queues = dev->online_queues - 1;
@@ -2057,8 +2056,12 @@ static int nvme_dev_add(struct nvme_dev *dev)
 		dev->tagset.flags = BLK_MQ_F_SHOULD_MERGE;
 		dev->tagset.driver_data = dev;
 
-		if (blk_mq_alloc_tag_set(&dev->tagset))
-			return 0;
+		ret = blk_mq_alloc_tag_set(&dev->tagset);
+		if (ret) {
+			dev_warn(dev->ctrl.device,
+				"IO queues tagset allocation failed %d\n", ret);
+			return ret;
+		}
 		dev->ctrl.tagset = &dev->tagset;
 
 		nvme_dbbuf_set(dev);
@@ -2291,6 +2294,7 @@ static void nvme_reset_work(struct work_struct *work)
 		container_of(work, struct nvme_dev, ctrl.reset_work);
 	bool was_suspend = !!(dev->ctrl.ctrl_config & NVME_CC_SHN_NORMAL);
 	int result = -ENODEV;
+	enum nvme_ctrl_state new_state = NVME_CTRL_LIVE;
 
 	if (WARN_ON(dev->ctrl.state != NVME_CTRL_RESETTING))
 		goto out;
@@ -2354,15 +2358,23 @@ static void nvme_reset_work(struct work_struct *work)
 		dev_warn(dev->ctrl.device, "IO queues not created\n");
 		nvme_kill_queues(&dev->ctrl);
 		nvme_remove_namespaces(&dev->ctrl);
+		new_state = NVME_CTRL_ADMIN_ONLY;
 	} else {
 		nvme_start_queues(&dev->ctrl);
 		nvme_wait_freeze(&dev->ctrl);
-		nvme_dev_add(dev);
+		/* hit this only when allocate tagset fails */
+		if (nvme_dev_add(dev))
+			new_state = NVME_CTRL_ADMIN_ONLY;
 		nvme_unfreeze(&dev->ctrl);
 	}
 
-	if (!nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_LIVE)) {
-		dev_warn(dev->ctrl.device, "failed to mark controller live\n");
+	/*
+	 * If only admin queue live, keep it to do further investigation or
+	 * recovery.
+	 */
+	if (!nvme_change_ctrl_state(&dev->ctrl, new_state)) {
+		dev_warn(dev->ctrl.device,
+			"failed to mark controller state %d\n", new_state);
 		goto out;
 	}
 
@@ -2498,10 +2510,10 @@ static int nvme_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (result)
 		goto release_pools;
 
-	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_RESETTING);
 	dev_info(dev->ctrl.device, "pci function %s\n", dev_name(&pdev->dev));
 
-	queue_work(nvme_wq, &dev->ctrl.reset_work);
+	nvme_reset_ctrl(&dev->ctrl);
+
 	return 0;
 
  release_pools:

drivers/nvme/host/rdma.c

@@ -2006,6 +2006,7 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
 
 static struct nvmf_transport_ops nvme_rdma_transport = {
 	.name		= "rdma",
+	.module		= THIS_MODULE,
 	.required_opts	= NVMF_OPT_TRADDR,
 	.allowed_opts	= NVMF_OPT_TRSVCID | NVMF_OPT_RECONNECT_DELAY |
 			  NVMF_OPT_HOST_TRADDR | NVMF_OPT_CTRL_LOSS_TMO,

drivers/nvme/target/core.c

@@ -830,7 +830,7 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 		/* Don't accept keep-alive timeout for discovery controllers */
 		if (kato) {
 			status = NVME_SC_INVALID_FIELD | NVME_SC_DNR;
-			goto out_free_sqs;
+			goto out_remove_ida;
 		}
 
 		/*
@@ -860,6 +860,8 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 	*ctrlp = ctrl;
 	return 0;
 
+out_remove_ida:
+	ida_simple_remove(&cntlid_ida, ctrl->cntlid);
 out_free_sqs:
 	kfree(ctrl->sqs);
 out_free_cqs:
@@ -877,21 +879,22 @@ static void nvmet_ctrl_free(struct kref *ref)
 	struct nvmet_ctrl *ctrl = container_of(ref, struct nvmet_ctrl, ref);
 	struct nvmet_subsys *subsys = ctrl->subsys;
 
-	nvmet_stop_keep_alive_timer(ctrl);
-
 	mutex_lock(&subsys->lock);
 	list_del(&ctrl->subsys_entry);
 	mutex_unlock(&subsys->lock);
 
+	nvmet_stop_keep_alive_timer(ctrl);
+
 	flush_work(&ctrl->async_event_work);
 	cancel_work_sync(&ctrl->fatal_err_work);
 
 	ida_simple_remove(&cntlid_ida, ctrl->cntlid);
-	nvmet_subsys_put(subsys);
 
 	kfree(ctrl->sqs);
 	kfree(ctrl->cqs);
 	kfree(ctrl);
+
+	nvmet_subsys_put(subsys);
 }
 
 void nvmet_ctrl_put(struct nvmet_ctrl *ctrl)

drivers/nvme/target/fabrics-cmd.c

@@ -225,7 +225,7 @@ static void nvmet_execute_io_connect(struct nvmet_req *req)
 		goto out_ctrl_put;
 	}
 
-	pr_info("adding queue %d to ctrl %d.\n", qid, ctrl->cntlid);
+	pr_debug("adding queue %d to ctrl %d.\n", qid, ctrl->cntlid);
 
 out:
 	kfree(d);

drivers/nvme/target/fc.c

@@ -2490,14 +2490,8 @@ nvmet_fc_add_port(struct nvmet_port *port)
 	list_for_each_entry(tgtport, &nvmet_fc_target_list, tgt_list) {
 		if ((tgtport->fc_target_port.node_name == traddr.nn) &&
 		    (tgtport->fc_target_port.port_name == traddr.pn)) {
-			/* a FC port can only be 1 nvmet port id */
-			if (!tgtport->port) {
-				tgtport->port = port;
-				port->priv = tgtport;
-				nvmet_fc_tgtport_get(tgtport);
-				ret = 0;
-			} else
-				ret = -EALREADY;
+			tgtport->port = port;
+			ret = 0;
 			break;
 		}
 	}
@@ -2508,19 +2502,7 @@ nvmet_fc_add_port(struct nvmet_port *port)
 static void
 nvmet_fc_remove_port(struct nvmet_port *port)
 {
-	struct nvmet_fc_tgtport *tgtport = port->priv;
-	unsigned long flags;
-	bool matched = false;
-
-	spin_lock_irqsave(&nvmet_fc_tgtlock, flags);
-	if (tgtport->port == port) {
-		matched = true;
-		tgtport->port = NULL;
-	}
-	spin_unlock_irqrestore(&nvmet_fc_tgtlock, flags);
-
-	if (matched)
-		nvmet_fc_tgtport_put(tgtport);
+	/* nothing to do */
 }
 
 static struct nvmet_fabrics_ops nvmet_fc_tgt_fcp_ops = {

drivers/nvme/target/loop.c

@@ -686,6 +686,7 @@ static struct nvmet_fabrics_ops nvme_loop_ops = {
 
 static struct nvmf_transport_ops nvme_loop_transport = {
 	.name		= "loop",
+	.module		= THIS_MODULE,
 	.create_ctrl	= nvme_loop_create_ctrl,
 };
 

drivers/nvme/target/rdma.c

@@ -921,7 +921,7 @@ static void nvmet_rdma_destroy_queue_ib(struct nvmet_rdma_queue *queue)
 
 static void nvmet_rdma_free_queue(struct nvmet_rdma_queue *queue)
 {
-	pr_info("freeing queue %d\n", queue->idx);
+	pr_debug("freeing queue %d\n", queue->idx);
 
 	nvmet_sq_destroy(&queue->nvme_sq);
 
@@ -1503,25 +1503,9 @@ static int __init nvmet_rdma_init(void)
 
 static void __exit nvmet_rdma_exit(void)
 {
-	struct nvmet_rdma_queue *queue;
-
 	nvmet_unregister_transport(&nvmet_rdma_ops);
-
-	flush_scheduled_work();
-
-	mutex_lock(&nvmet_rdma_queue_mutex);
-	while ((queue = list_first_entry_or_null(&nvmet_rdma_queue_list,
-			struct nvmet_rdma_queue, queue_list))) {
-		list_del_init(&queue->queue_list);
-
-		mutex_unlock(&nvmet_rdma_queue_mutex);
-		__nvmet_rdma_queue_disconnect(queue);
-		mutex_lock(&nvmet_rdma_queue_mutex);
-	}
-	mutex_unlock(&nvmet_rdma_queue_mutex);
-
-	flush_scheduled_work();
 	ib_unregister_client(&nvmet_rdma_ib_client);
+	WARN_ON_ONCE(!list_empty(&nvmet_rdma_queue_list));
 	ida_destroy(&nvmet_rdma_queue_ida);
 }