Merge tag 'nvme-5.16-2021-10-21' of git://git.infradead.org/nvme into for-5.16/drivers

Pull NVMe updates from Christoph:

"nvme updates for Linux 5.16

 - fix a multipath partition scanning deadlock (Hannes Reinecke)
 - generate uevent once a multipath namespace is operational again
   (Hannes Reinecke)
 - support unique discovery controller NQNs (Hannes Reinecke)
 - fix use-after-free when a port is removed (Israel Rukshin)
 - clear shadow doorbell memory on resets (Keith Busch)
 - use struct_size (Len Baker)
 - add error handling support for add_disk (Luis Chamberlain)
 - limit the maximal queue size for RDMA controllers (Max Gurtovoy)
 - use a few more symbolic names (Max Gurtovoy)
 - fix error code in nvme_rdma_setup_ctrl (Max Gurtovoy)
 - add support for ->map_queues on FC (Saurav Kashyap)"

* tag 'nvme-5.16-2021-10-21' of git://git.infradead.org/nvme: (23 commits)
  nvmet: use struct_size over open coded arithmetic
  nvme: drop scan_lock and always kick requeue list when removing namespaces
  nvme-pci: clear shadow doorbell memory on resets
  nvme-rdma: fix error code in nvme_rdma_setup_ctrl
  nvme-multipath: add error handling support for add_disk()
  nvmet: use macro definitions for setting cmic value
  nvmet: use macro definition for setting nmic value
  nvme: display correct subsystem NQN
  nvme: Add connect option 'discovery'
  nvme: expose subsystem type in sysfs attribute 'subsystype'
  nvmet: set 'CNTRLTYPE' in the identify controller data
  nvmet: add nvmet_is_disc_subsys() helper
  nvme: add CNTRLTYPE definitions for 'identify controller'
  nvmet: make discovery NQN configurable
  nvmet-rdma: implement get_max_queue_size controller op
  nvmet: add get_max_queue_size op for controllers
  nvme-rdma: limit the maximal queue size for RDMA controllers
  nvmet-tcp: fix use-after-free when a port is removed
  nvmet-rdma: fix use-after-free when a port is removed
  nvmet: fix use-after-free when a port is removed
  ...

drivers/nvme/host/core.c

@@ -222,7 +222,7 @@ int nvme_reset_ctrl_sync(struct nvme_ctrl *ctrl)
 static void nvme_do_delete_ctrl(struct nvme_ctrl *ctrl)
 {
 	dev_info(ctrl->device,
-		 "Removing ctrl: NQN \"%s\"\n", ctrl->opts->subsysnqn);
+		 "Removing ctrl: NQN \"%s\"\n", nvmf_ctrl_subsysnqn(ctrl));
 
 	flush_work(&ctrl->reset_work);
 	nvme_stop_ctrl(ctrl);
@@ -2620,6 +2620,24 @@ static ssize_t nvme_subsys_show_nqn(struct device *dev,
 }
 static SUBSYS_ATTR_RO(subsysnqn, S_IRUGO, nvme_subsys_show_nqn);
 
+static ssize_t nvme_subsys_show_type(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct nvme_subsystem *subsys =
+		container_of(dev, struct nvme_subsystem, dev);
+
+	switch (subsys->subtype) {
+	case NVME_NQN_DISC:
+		return sysfs_emit(buf, "discovery\n");
+	case NVME_NQN_NVME:
+		return sysfs_emit(buf, "nvm\n");
+	default:
+		return sysfs_emit(buf, "reserved\n");
+	}
+}
+static SUBSYS_ATTR_RO(subsystype, S_IRUGO, nvme_subsys_show_type);
+
 #define nvme_subsys_show_str_function(field)				\
 static ssize_t subsys_##field##_show(struct device *dev,		\
 			    struct device_attribute *attr, char *buf)	\
@@ -2640,6 +2658,7 @@ static struct attribute *nvme_subsys_attrs[] = {
 	&subsys_attr_serial.attr,
 	&subsys_attr_firmware_rev.attr,
 	&subsys_attr_subsysnqn.attr,
+	&subsys_attr_subsystype.attr,
 #ifdef CONFIG_NVME_MULTIPATH
 	&subsys_attr_iopolicy.attr,
 #endif
@@ -2710,6 +2729,21 @@ static int nvme_init_subsystem(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id)
 	memcpy(subsys->firmware_rev, id->fr, sizeof(subsys->firmware_rev));
 	subsys->vendor_id = le16_to_cpu(id->vid);
 	subsys->cmic = id->cmic;
+
+	/* Versions prior to 1.4 don't necessarily report a valid type */
+	if (id->cntrltype == NVME_CTRL_DISC ||
+	    !strcmp(subsys->subnqn, NVME_DISC_SUBSYS_NAME))
+		subsys->subtype = NVME_NQN_DISC;
+	else
+		subsys->subtype = NVME_NQN_NVME;
+
+	if (nvme_discovery_ctrl(ctrl) && subsys->subtype != NVME_NQN_DISC) {
+		dev_err(ctrl->device,
+			"Subsystem %s is not a discovery controller",
+			subsys->subnqn);
+		kfree(subsys);
+		return -EINVAL;
+	}
 	subsys->awupf = le16_to_cpu(id->awupf);
 #ifdef CONFIG_NVME_MULTIPATH
 	subsys->iopolicy = NVME_IOPOLICY_NUMA;

drivers/nvme/host/fabrics.c

@@ -548,6 +548,7 @@ static const match_table_t opt_tokens = {
 	{ NVMF_OPT_NR_POLL_QUEUES,	"nr_poll_queues=%d"	},
 	{ NVMF_OPT_TOS,			"tos=%d"		},
 	{ NVMF_OPT_FAIL_FAST_TMO,	"fast_io_fail_tmo=%d"	},
+	{ NVMF_OPT_DISCOVERY,		"discovery"		},
 	{ NVMF_OPT_ERR,			NULL			}
 };
 
@@ -823,6 +824,9 @@ static int nvmf_parse_options(struct nvmf_ctrl_options *opts,
 			}
 			opts->tos = token;
 			break;
+		case NVMF_OPT_DISCOVERY:
+			opts->discovery_nqn = true;
+			break;
 		default:
 			pr_warn("unknown parameter or missing value '%s' in ctrl creation request\n",
 				p);
@@ -949,7 +953,7 @@ EXPORT_SYMBOL_GPL(nvmf_free_options);
 #define NVMF_ALLOWED_OPTS	(NVMF_OPT_QUEUE_SIZE | NVMF_OPT_NR_IO_QUEUES | \
 				 NVMF_OPT_KATO | NVMF_OPT_HOSTNQN | \
 				 NVMF_OPT_HOST_ID | NVMF_OPT_DUP_CONNECT |\
-				 NVMF_OPT_DISABLE_SQFLOW |\
+				 NVMF_OPT_DISABLE_SQFLOW | NVMF_OPT_DISCOVERY |\
 				 NVMF_OPT_FAIL_FAST_TMO)
 
 static struct nvme_ctrl *

drivers/nvme/host/fabrics.h

@@ -67,6 +67,7 @@ enum {
 	NVMF_OPT_TOS		= 1 << 19,
 	NVMF_OPT_FAIL_FAST_TMO	= 1 << 20,
 	NVMF_OPT_HOST_IFACE	= 1 << 21,
+	NVMF_OPT_DISCOVERY	= 1 << 22,
 };
 
 /**
@@ -178,6 +179,13 @@ nvmf_ctlr_matches_baseopts(struct nvme_ctrl *ctrl,
 	return true;
 }
 
+static inline char *nvmf_ctrl_subsysnqn(struct nvme_ctrl *ctrl)
+{
+	if (!ctrl->subsys)
+		return ctrl->opts->subsysnqn;
+	return ctrl->subsys->subnqn;
+}
+
 int nvmf_reg_read32(struct nvme_ctrl *ctrl, u32 off, u32 *val);
 int nvmf_reg_read64(struct nvme_ctrl *ctrl, u32 off, u64 *val);
 int nvmf_reg_write32(struct nvme_ctrl *ctrl, u32 off, u32 val);

drivers/nvme/host/fc.c

@@ -16,6 +16,7 @@
 #include <linux/nvme-fc.h>
 #include "fc.h"
 #include <scsi/scsi_transport_fc.h>
+#include <linux/blk-mq-pci.h>
 
 /* *************************** Data Structures/Defines ****************** */
 
@@ -2841,6 +2842,28 @@ nvme_fc_complete_rq(struct request *rq)
 	nvme_fc_ctrl_put(ctrl);
 }
 
+static int nvme_fc_map_queues(struct blk_mq_tag_set *set)
+{
+	struct nvme_fc_ctrl *ctrl = set->driver_data;
+	int i;
+
+	for (i = 0; i < set->nr_maps; i++) {
+		struct blk_mq_queue_map *map = &set->map[i];
+
+		if (!map->nr_queues) {
+			WARN_ON(i == HCTX_TYPE_DEFAULT);
+			continue;
+		}
+
+		/* Call LLDD map queue functionality if defined */
+		if (ctrl->lport->ops->map_queues)
+			ctrl->lport->ops->map_queues(&ctrl->lport->localport,
+						     map);
+		else
+			blk_mq_map_queues(map);
+	}
+	return 0;
+}
 
 static const struct blk_mq_ops nvme_fc_mq_ops = {
 	.queue_rq	= nvme_fc_queue_rq,
@@ -2849,6 +2872,7 @@ static const struct blk_mq_ops nvme_fc_mq_ops = {
 	.exit_request	= nvme_fc_exit_request,
 	.init_hctx	= nvme_fc_init_hctx,
 	.timeout	= nvme_fc_timeout,
+	.map_queues	= nvme_fc_map_queues,
 };
 
 static int
@@ -3572,7 +3596,7 @@ nvme_fc_init_ctrl(struct device *dev, struct nvmf_ctrl_options *opts,
 
 	dev_info(ctrl->ctrl.device,
 		"NVME-FC{%d}: new ctrl: NQN \"%s\"\n",
-		ctrl->cnum, ctrl->ctrl.opts->subsysnqn);
+		ctrl->cnum, nvmf_ctrl_subsysnqn(&ctrl->ctrl));
 
 	return &ctrl->ctrl;
 

drivers/nvme/host/multipath.c

@@ -105,8 +105,11 @@ void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl)
 
 	down_read(&ctrl->namespaces_rwsem);
 	list_for_each_entry(ns, &ctrl->namespaces, list) {
-		if (ns->head->disk)
-			kblockd_schedule_work(&ns->head->requeue_work);
+		if (!ns->head->disk)
+			continue;
+		kblockd_schedule_work(&ns->head->requeue_work);
+		if (ctrl->state == NVME_CTRL_LIVE)
+			disk_uevent(ns->head->disk, KOBJ_CHANGE);
 	}
 	up_read(&ctrl->namespaces_rwsem);
 }
@@ -143,13 +146,12 @@ void nvme_mpath_clear_ctrl_paths(struct nvme_ctrl *ctrl)
 {
 	struct nvme_ns *ns;
 
-	mutex_lock(&ctrl->scan_lock);
 	down_read(&ctrl->namespaces_rwsem);
-	list_for_each_entry(ns, &ctrl->namespaces, list)
-		if (nvme_mpath_clear_current_path(ns))
-			kblockd_schedule_work(&ns->head->requeue_work);
+	list_for_each_entry(ns, &ctrl->namespaces, list) {
+		nvme_mpath_clear_current_path(ns);
+		kblockd_schedule_work(&ns->head->requeue_work);
+	}
 	up_read(&ctrl->namespaces_rwsem);
-	mutex_unlock(&ctrl->scan_lock);
 }
 
 void nvme_mpath_revalidate_paths(struct nvme_ns *ns)
@@ -506,13 +508,23 @@ int nvme_mpath_alloc_disk(struct nvme_ctrl *ctrl, struct nvme_ns_head *head)
 static void nvme_mpath_set_live(struct nvme_ns *ns)
 {
 	struct nvme_ns_head *head = ns->head;
+	int rc;
 
 	if (!head->disk)
 		return;
 
+	/*
+	 * test_and_set_bit() is used because it is protecting against two nvme
+	 * paths simultaneously calling device_add_disk() on the same namespace
+	 * head.
+	 */
 	if (!test_and_set_bit(NVME_NSHEAD_DISK_LIVE, &head->flags)) {
-		device_add_disk(&head->subsys->dev, head->disk,
-				nvme_ns_id_attr_groups);
+		rc = device_add_disk(&head->subsys->dev, head->disk,
+				     nvme_ns_id_attr_groups);
+		if (rc) {
+			clear_bit(NVME_NSHEAD_DISK_LIVE, &ns->flags);
+			return;
+		}
 		nvme_add_ns_head_cdev(head);
 	}
 

drivers/nvme/host/nvme.h

@@ -372,6 +372,7 @@ struct nvme_subsystem {
 	char			model[40];
 	char			firmware_rev[8];
 	u8			cmic;
+	enum nvme_subsys_type	subtype;
 	u16			vendor_id;
 	u16			awupf;	/* 0's based awupf value. */
 	struct ida		ns_ida;

drivers/nvme/host/pci.c

@@ -245,8 +245,15 @@ static int nvme_dbbuf_dma_alloc(struct nvme_dev *dev)
 {
 	unsigned int mem_size = nvme_dbbuf_size(dev);
 
-	if (dev->dbbuf_dbs)
+	if (dev->dbbuf_dbs) {
+		/*
+		 * Clear the dbbuf memory so the driver doesn't observe stale
+		 * values from the previous instantiation.
+		 */
+		memset(dev->dbbuf_dbs, 0, mem_size);
+		memset(dev->dbbuf_eis, 0, mem_size);
 		return 0;
+	}
 
 	dev->dbbuf_dbs = dma_alloc_coherent(dev->dev, mem_size,
 					    &dev->dbbuf_dbs_dma_addr,

drivers/nvme/host/rdma.c

@@ -1096,11 +1096,13 @@ static int nvme_rdma_setup_ctrl(struct nvme_rdma_ctrl *ctrl, bool new)
 		return ret;
 
 	if (ctrl->ctrl.icdoff) {
+		ret = -EOPNOTSUPP;
 		dev_err(ctrl->ctrl.device, "icdoff is not supported!\n");
 		goto destroy_admin;
 	}
 
 	if (!(ctrl->ctrl.sgls & (1 << 2))) {
+		ret = -EOPNOTSUPP;
 		dev_err(ctrl->ctrl.device,
 			"Mandatory keyed sgls are not supported!\n");
 		goto destroy_admin;
@@ -1112,6 +1114,13 @@ static int nvme_rdma_setup_ctrl(struct nvme_rdma_ctrl *ctrl, bool new)
 			ctrl->ctrl.opts->queue_size, ctrl->ctrl.sqsize + 1);
 	}
 
+	if (ctrl->ctrl.sqsize + 1 > NVME_RDMA_MAX_QUEUE_SIZE) {
+		dev_warn(ctrl->ctrl.device,
+			"ctrl sqsize %u > max queue size %u, clamping down\n",
+			ctrl->ctrl.sqsize + 1, NVME_RDMA_MAX_QUEUE_SIZE);
+		ctrl->ctrl.sqsize = NVME_RDMA_MAX_QUEUE_SIZE - 1;
+	}
+
 	if (ctrl->ctrl.sqsize + 1 > ctrl->ctrl.maxcmd) {
 		dev_warn(ctrl->ctrl.device,
 			"sqsize %u > ctrl maxcmd %u, clamping down\n",
@@ -2386,7 +2395,7 @@ static struct nvme_ctrl *nvme_rdma_create_ctrl(struct device *dev,
 		goto out_uninit_ctrl;
 
 	dev_info(ctrl->ctrl.device, "new ctrl: NQN \"%s\", addr %pISpcs\n",
-		ctrl->ctrl.opts->subsysnqn, &ctrl->addr);
+		nvmf_ctrl_subsysnqn(&ctrl->ctrl), &ctrl->addr);
 
 	mutex_lock(&nvme_rdma_ctrl_mutex);
 	list_add_tail(&ctrl->list, &nvme_rdma_ctrl_list);

drivers/nvme/host/tcp.c

@@ -2582,7 +2582,7 @@ static struct nvme_ctrl *nvme_tcp_create_ctrl(struct device *dev,
 		goto out_uninit_ctrl;
 
 	dev_info(ctrl->ctrl.device, "new ctrl: NQN \"%s\", addr %pISp\n",
-		ctrl->ctrl.opts->subsysnqn, &ctrl->addr);
+		nvmf_ctrl_subsysnqn(&ctrl->ctrl), &ctrl->addr);
 
 	mutex_lock(&nvme_tcp_ctrl_mutex);
 	list_add_tail(&ctrl->list, &nvme_tcp_ctrl_list);

drivers/nvme/target/admin-cmd.c

@@ -278,8 +278,8 @@ static void nvmet_execute_get_log_page_ana(struct nvmet_req *req)
 	u16 status;
 
 	status = NVME_SC_INTERNAL;
-	desc = kmalloc(sizeof(struct nvme_ana_group_desc) +
-			NVMET_MAX_NAMESPACES * sizeof(__le32), GFP_KERNEL);
+	desc = kmalloc(struct_size(desc, nsids, NVMET_MAX_NAMESPACES),
+		       GFP_KERNEL);
 	if (!desc)
 		goto out;
 
@@ -374,13 +374,19 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
 
 	id->rab = 6;
 
+	if (nvmet_is_disc_subsys(ctrl->subsys))
+		id->cntrltype = NVME_CTRL_DISC;
+	else
+		id->cntrltype = NVME_CTRL_IO;
+
 	/*
 	 * XXX: figure out how we can assign a IEEE OUI, but until then
 	 * the safest is to leave it as zeroes.
 	 */
 
 	/* we support multiple ports, multiples hosts and ANA: */
-	id->cmic = (1 << 0) | (1 << 1) | (1 << 3);
+	id->cmic = NVME_CTRL_CMIC_MULTI_PORT | NVME_CTRL_CMIC_MULTI_CTRL |
+		NVME_CTRL_CMIC_ANA;
 
 	/* Limit MDTS according to transport capability */
 	if (ctrl->ops->get_mdts)
@@ -536,7 +542,7 @@ static void nvmet_execute_identify_ns(struct nvmet_req *req)
 	 * Our namespace might always be shared.  Not just with other
 	 * controllers, but also with any other user of the block device.
 	 */
-	id->nmic = (1 << 0);
+	id->nmic = NVME_NS_NMIC_SHARED;
 	id->anagrpid = cpu_to_le32(req->ns->anagrpid);
 
 	memcpy(&id->nguid, &req->ns->nguid, sizeof(id->nguid));
@@ -1008,7 +1014,7 @@ u16 nvmet_parse_admin_cmd(struct nvmet_req *req)
 
 	if (nvme_is_fabrics(cmd))
 		return nvmet_parse_fabrics_cmd(req);
-	if (nvmet_req_subsys(req)->type == NVME_NQN_DISC)
+	if (nvmet_is_disc_subsys(nvmet_req_subsys(req)))
 		return nvmet_parse_discovery_cmd(req);
 
 	ret = nvmet_check_ctrl_status(req);

drivers/nvme/target/configfs.c

@@ -1233,6 +1233,44 @@ static ssize_t nvmet_subsys_attr_model_store(struct config_item *item,
 }
 CONFIGFS_ATTR(nvmet_subsys_, attr_model);
 
+static ssize_t nvmet_subsys_attr_discovery_nqn_show(struct config_item *item,
+			char *page)
+{
+	return snprintf(page, PAGE_SIZE, "%s\n",
+			nvmet_disc_subsys->subsysnqn);
+}
+
+static ssize_t nvmet_subsys_attr_discovery_nqn_store(struct config_item *item,
+			const char *page, size_t count)
+{
+	struct nvmet_subsys *subsys = to_subsys(item);
+	char *subsysnqn;
+	int len;
+
+	len = strcspn(page, "\n");
+	if (!len)
+		return -EINVAL;
+
+	subsysnqn = kmemdup_nul(page, len, GFP_KERNEL);
+	if (!subsysnqn)
+		return -ENOMEM;
+
+	/*
+	 * The discovery NQN must be different from subsystem NQN.
+	 */
+	if (!strcmp(subsysnqn, subsys->subsysnqn)) {
+		kfree(subsysnqn);
+		return -EBUSY;
+	}
+	down_write(&nvmet_config_sem);
+	kfree(nvmet_disc_subsys->subsysnqn);
+	nvmet_disc_subsys->subsysnqn = subsysnqn;
+	up_write(&nvmet_config_sem);
+
+	return count;
+}
+CONFIGFS_ATTR(nvmet_subsys_, attr_discovery_nqn);
+
 #ifdef CONFIG_BLK_DEV_INTEGRITY
 static ssize_t nvmet_subsys_attr_pi_enable_show(struct config_item *item,
 						char *page)
@@ -1262,6 +1300,7 @@ static struct configfs_attribute *nvmet_subsys_attrs[] = {
 	&nvmet_subsys_attr_attr_cntlid_min,
 	&nvmet_subsys_attr_attr_cntlid_max,
 	&nvmet_subsys_attr_attr_model,
+	&nvmet_subsys_attr_attr_discovery_nqn,
 #ifdef CONFIG_BLK_DEV_INTEGRITY
 	&nvmet_subsys_attr_attr_pi_enable,
 #endif
@@ -1553,6 +1592,8 @@ static void nvmet_port_release(struct config_item *item)
 {
 	struct nvmet_port *port = to_nvmet_port(item);
 
+	/* Let inflight controllers teardown complete */
+	flush_scheduled_work();
 	list_del(&port->global_entry);
 
 	kfree(port->ana_state);

drivers/nvme/target/core.c

@@ -1140,7 +1140,7 @@ static void nvmet_start_ctrl(struct nvmet_ctrl *ctrl)
 	 * should verify iosqes,iocqes are zeroed, however that
 	 * would break backwards compatibility, so don't enforce it.
 	 */
-	if (ctrl->subsys->type != NVME_NQN_DISC &&
+	if (!nvmet_is_disc_subsys(ctrl->subsys) &&
 	    (nvmet_cc_iosqes(ctrl->cc) != NVME_NVM_IOSQES ||
 	     nvmet_cc_iocqes(ctrl->cc) != NVME_NVM_IOCQES)) {
 		ctrl->csts = NVME_CSTS_CFS;
@@ -1205,7 +1205,10 @@ static void nvmet_init_cap(struct nvmet_ctrl *ctrl)
 	/* CC.EN timeout in 500msec units: */
 	ctrl->cap |= (15ULL << 24);
 	/* maximum queue entries supported: */
-	ctrl->cap |= NVMET_QUEUE_SIZE - 1;
+	if (ctrl->ops->get_max_queue_size)
+		ctrl->cap |= ctrl->ops->get_max_queue_size(ctrl) - 1;
+	else
+		ctrl->cap |= NVMET_QUEUE_SIZE - 1;
 
 	if (nvmet_is_passthru_subsys(ctrl->subsys))
 		nvmet_passthrough_override_cap(ctrl);
@@ -1278,7 +1281,7 @@ bool nvmet_host_allowed(struct nvmet_subsys *subsys, const char *hostnqn)
 	if (subsys->allow_any_host)
 		return true;
 
-	if (subsys->type == NVME_NQN_DISC) /* allow all access to disc subsys */
+	if (nvmet_is_disc_subsys(subsys)) /* allow all access to disc subsys */
 		return true;
 
 	list_for_each_entry(p, &subsys->hosts, entry) {
@@ -1367,6 +1370,7 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 	mutex_init(&ctrl->lock);
 
 	ctrl->port = req->port;
+	ctrl->ops = req->ops;
 
 	INIT_WORK(&ctrl->async_event_work, nvmet_async_event_work);
 	INIT_LIST_HEAD(&ctrl->async_events);
@@ -1405,13 +1409,11 @@ u16 nvmet_alloc_ctrl(const char *subsysnqn, const char *hostnqn,
 	}
 	ctrl->cntlid = ret;
 
-	ctrl->ops = req->ops;
-
 	/*
 	 * Discovery controllers may use some arbitrary high value
 	 * in order to cleanup stale discovery sessions
 	 */
-	if ((ctrl->subsys->type == NVME_NQN_DISC) && !kato)
+	if (nvmet_is_disc_subsys(ctrl->subsys) && !kato)
 		kato = NVMET_DISC_KATO_MS;
 
 	/* keep-alive timeout in seconds */
@@ -1491,7 +1493,8 @@ static struct nvmet_subsys *nvmet_find_get_subsys(struct nvmet_port *port,
 	if (!port)
 		return NULL;
 
-	if (!strcmp(NVME_DISC_SUBSYS_NAME, subsysnqn)) {
+	if (!strcmp(NVME_DISC_SUBSYS_NAME, subsysnqn) ||
+	    !strcmp(nvmet_disc_subsys->subsysnqn, subsysnqn)) {
 		if (!kref_get_unless_zero(&nvmet_disc_subsys->ref))
 			return NULL;
 		return nvmet_disc_subsys;

drivers/nvme/target/discovery.c

@@ -268,6 +268,8 @@ static void nvmet_execute_disc_identify(struct nvmet_req *req)
 	memcpy_and_pad(id->fr, sizeof(id->fr),
 		       UTS_RELEASE, strlen(UTS_RELEASE), ' ');
 
+	id->cntrltype = NVME_CTRL_DISC;
+
 	/* no limit on data transfer sizes for now */
 	id->mdts = 0;
 	id->cntlid = cpu_to_le16(ctrl->cntlid);

drivers/nvme/target/fabrics-cmd.c

@@ -221,7 +221,8 @@ static void nvmet_execute_admin_connect(struct nvmet_req *req)
 		goto out;
 	}
 
-	pr_info("creating controller %d for subsystem %s for NQN %s%s.\n",
+	pr_info("creating %s controller %d for subsystem %s for NQN %s%s.\n",
+		nvmet_is_disc_subsys(ctrl->subsys) ? "discovery" : "nvm",
 		ctrl->cntlid, ctrl->subsys->subsysnqn, ctrl->hostnqn,
 		ctrl->pi_support ? " T10-PI is enabled" : "");
 	req->cqe->result.u16 = cpu_to_le16(ctrl->cntlid);

drivers/nvme/target/nvmet.h

@@ -309,6 +309,7 @@ struct nvmet_fabrics_ops {
 	u16 (*install_queue)(struct nvmet_sq *nvme_sq);
 	void (*discovery_chg)(struct nvmet_port *port);
 	u8 (*get_mdts)(const struct nvmet_ctrl *ctrl);
+	u16 (*get_max_queue_size)(const struct nvmet_ctrl *ctrl);
 };
 
 #define NVMET_MAX_INLINE_BIOVEC	8
@@ -576,6 +577,11 @@ static inline struct nvmet_subsys *nvmet_req_subsys(struct nvmet_req *req)
 	return req->sq->ctrl->subsys;
 }
 
+static inline bool nvmet_is_disc_subsys(struct nvmet_subsys *subsys)
+{
+    return subsys->type == NVME_NQN_DISC;
+}
+
 #ifdef CONFIG_NVME_TARGET_PASSTHRU
 void nvmet_passthru_subsys_free(struct nvmet_subsys *subsys);
 int nvmet_passthru_ctrl_enable(struct nvmet_subsys *subsys);

drivers/nvme/target/rdma.c

@@ -1819,12 +1819,36 @@ static void nvmet_rdma_delete_ctrl(struct nvmet_ctrl *ctrl)
 	mutex_unlock(&nvmet_rdma_queue_mutex);
 }
 
+static void nvmet_rdma_destroy_port_queues(struct nvmet_rdma_port *port)
+{
+	struct nvmet_rdma_queue *queue, *tmp;
+	struct nvmet_port *nport = port->nport;
+
+	mutex_lock(&nvmet_rdma_queue_mutex);
+	list_for_each_entry_safe(queue, tmp, &nvmet_rdma_queue_list,
+				 queue_list) {
+		if (queue->port != nport)
+			continue;
+
+		list_del_init(&queue->queue_list);
+		__nvmet_rdma_queue_disconnect(queue);
+	}
+	mutex_unlock(&nvmet_rdma_queue_mutex);
+}
+
 static void nvmet_rdma_disable_port(struct nvmet_rdma_port *port)
 {
 	struct rdma_cm_id *cm_id = xchg(&port->cm_id, NULL);
 
 	if (cm_id)
 		rdma_destroy_id(cm_id);
+
+	/*
+	 * Destroy the remaining queues, which are not belong to any
+	 * controller yet. Do it here after the RDMA-CM was destroyed
+	 * guarantees that no new queue will be created.
+	 */
+	nvmet_rdma_destroy_port_queues(port);
 }
 
 static int nvmet_rdma_enable_port(struct nvmet_rdma_port *port)
@@ -1976,6 +2000,11 @@ static u8 nvmet_rdma_get_mdts(const struct nvmet_ctrl *ctrl)
 	return NVMET_RDMA_MAX_MDTS;
 }
 
+static u16 nvmet_rdma_get_max_queue_size(const struct nvmet_ctrl *ctrl)
+{
+	return NVME_RDMA_MAX_QUEUE_SIZE;
+}
+
 static const struct nvmet_fabrics_ops nvmet_rdma_ops = {
 	.owner			= THIS_MODULE,
 	.type			= NVMF_TRTYPE_RDMA,
@@ -1987,6 +2016,7 @@ static const struct nvmet_fabrics_ops nvmet_rdma_ops = {
 	.delete_ctrl		= nvmet_rdma_delete_ctrl,
 	.disc_traddr		= nvmet_rdma_disc_port_addr,
 	.get_mdts		= nvmet_rdma_get_mdts,
+	.get_max_queue_size	= nvmet_rdma_get_max_queue_size,
 };
 
 static void nvmet_rdma_remove_one(struct ib_device *ib_device, void *client_data)

drivers/nvme/target/tcp.c

@@ -1737,6 +1737,17 @@ static int nvmet_tcp_add_port(struct nvmet_port *nport)
 	return ret;
 }
 
+static void nvmet_tcp_destroy_port_queues(struct nvmet_tcp_port *port)
+{
+	struct nvmet_tcp_queue *queue;
+
+	mutex_lock(&nvmet_tcp_queue_mutex);
+	list_for_each_entry(queue, &nvmet_tcp_queue_list, queue_list)
+		if (queue->port == port)
+			kernel_sock_shutdown(queue->sock, SHUT_RDWR);
+	mutex_unlock(&nvmet_tcp_queue_mutex);
+}
+
 static void nvmet_tcp_remove_port(struct nvmet_port *nport)
 {
 	struct nvmet_tcp_port *port = nport->priv;
@@ -1746,6 +1757,11 @@ static void nvmet_tcp_remove_port(struct nvmet_port *nport)
 	port->sock->sk->sk_user_data = NULL;
 	write_unlock_bh(&port->sock->sk->sk_callback_lock);
 	cancel_work_sync(&port->accept_work);
+	/*
+	 * Destroy the remaining queues, which are not belong to any
+	 * controller yet.
+	 */
+	nvmet_tcp_destroy_port_queues(port);
 
 	sock_release(port->sock);
 	kfree(port);

drivers/scsi/qla2xxx/qla_nvme.c

@@ -8,6 +8,8 @@
 #include <linux/delay.h>
 #include <linux/nvme.h>
 #include <linux/nvme-fc.h>
+#include <linux/blk-mq-pci.h>
+#include <linux/blk-mq.h>
 
 static struct nvme_fc_port_template qla_nvme_fc_transport;
 
@@ -642,6 +644,18 @@ static int qla_nvme_post_cmd(struct nvme_fc_local_port *lport,
 	return rval;
 }
 
+static void qla_nvme_map_queues(struct nvme_fc_local_port *lport,
+		struct blk_mq_queue_map *map)
+{
+	struct scsi_qla_host *vha = lport->private;
+	int rc;
+
+	rc = blk_mq_pci_map_queues(map, vha->hw->pdev, vha->irq_offset);
+	if (rc)
+		ql_log(ql_log_warn, vha, 0x21de,
+		       "pci map queue failed 0x%x", rc);
+}
+
 static void qla_nvme_localport_delete(struct nvme_fc_local_port *lport)
 {
 	struct scsi_qla_host *vha = lport->private;
@@ -676,6 +690,7 @@ static struct nvme_fc_port_template qla_nvme_fc_transport = {
 	.ls_abort	= qla_nvme_ls_abort,
 	.fcp_io		= qla_nvme_post_cmd,
 	.fcp_abort	= qla_nvme_fcp_abort,
+	.map_queues	= qla_nvme_map_queues,
 	.max_hw_queues  = 8,
 	.max_sgl_segments = 1024,
 	.max_dif_sgl_segments = 64,

include/linux/nvme-fc-driver.h

@@ -7,6 +7,7 @@
 #define _NVME_FC_DRIVER_H 1
 
 #include <linux/scatterlist.h>
+#include <linux/blk-mq.h>
 
 
 /*
@@ -497,6 +498,8 @@ struct nvme_fc_port_template {
 	int	(*xmt_ls_rsp)(struct nvme_fc_local_port *localport,
 				struct nvme_fc_remote_port *rport,
 				struct nvmefc_ls_rsp *ls_rsp);
+	void	(*map_queues)(struct nvme_fc_local_port *localport,
+			      struct blk_mq_queue_map *map);
 
 	u32	max_hw_queues;
 	u16	max_sgl_segments;
@@ -779,6 +782,10 @@ struct nvmet_fc_target_port {
  *       LS received.
  *       Entrypoint is Mandatory.
  *
+ * @map_queues: This functions lets the driver expose the queue mapping
+ *	 to the block layer.
+ *       Entrypoint is Optional.
+ *
  * @fcp_op:  Called to perform a data transfer or transmit a response.
  *       The nvmefc_tgt_fcp_req structure is the same LLDD-supplied
  *       exchange structure specified in the nvmet_fc_rcv_fcp_req() call

include/linux/nvme-rdma.h

@@ -6,6 +6,8 @@
 #ifndef _LINUX_NVME_RDMA_H
 #define _LINUX_NVME_RDMA_H
 
+#define NVME_RDMA_MAX_QUEUE_SIZE	128
+
 enum nvme_rdma_cm_fmt {
 	NVME_RDMA_CM_FMT_1_0 = 0x0,
 };

include/linux/nvme.h

@@ -31,6 +31,12 @@ enum nvme_subsys_type {
 	NVME_NQN_NVME	= 2,		/* NVME type target subsystem */
 };
 
+enum nvme_ctrl_type {
+	NVME_CTRL_IO	= 1,		/* I/O controller */
+	NVME_CTRL_DISC	= 2,		/* Discovery controller */
+	NVME_CTRL_ADMIN	= 3,		/* Administrative controller */
+};
+
 /* Address Family codes for Discovery Log Page entry ADRFAM field */
 enum {
 	NVMF_ADDR_FAMILY_PCI	= 0,	/* PCIe */
@@ -244,7 +250,9 @@ struct nvme_id_ctrl {
 	__le32			rtd3e;
 	__le32			oaes;
 	__le32			ctratt;
-	__u8			rsvd100[28];
+	__u8			rsvd100[11];
+	__u8			cntrltype;
+	__u8			fguid[16];
 	__le16			crdt1;
 	__le16			crdt2;
 	__le16			crdt3;
@@ -312,6 +320,7 @@ struct nvme_id_ctrl {
 };
 
 enum {
+	NVME_CTRL_CMIC_MULTI_PORT		= 1 << 0,
 	NVME_CTRL_CMIC_MULTI_CTRL		= 1 << 1,
 	NVME_CTRL_CMIC_ANA			= 1 << 3,
 	NVME_CTRL_ONCS_COMPARE			= 1 << 0,