nvme-multipath: Add visibility for round-robin io-policy

This patch helps add nvme native multipath visibility for round-robin
io-policy. It creates a "multipath" sysfs directory under head gendisk
device node directory and then from "multipath" directory it adds a link
to each namespace path device the head node refers.

For instance, if we have a shared namespace accessible from two different
controllers/paths then we create a soft link to each path device from head
disk node as shown below:

$ ls -l /sys/block/nvme1n1/multipath/
nvme1c1n1 -> ../../../../../pci052e:78/052e:78:00.0/nvme/nvme1/nvme1c1n1
nvme1c3n1 -> ../../../../../pci058e:78/058e:78:00.0/nvme/nvme3/nvme1c3n1

In the above example, nvme1n1 is head gendisk node created for a shared
namespace and the namespace is accessible from nvme1c1n1 and nvme1c3n1
paths.

For round-robin I/O policy, we could easily infer from the above output
that I/O workload targeted to nvme1n1 would toggle across paths nvme1c1n1
and nvme1c3n1.

Reviewed-by: Hannes Reinecke <[email protected]>
Signed-off-by: Nilay Shroff <[email protected]>
Signed-off-by: Keith Busch <[email protected]>

Author: shroffni

drivers/nvme/host/core.c

@@ -4020,6 +4020,9 @@ static void nvme_ns_remove(struct nvme_ns *ns)
 
 	if (!nvme_ns_head_multipath(ns->head))
 		nvme_cdev_del(&ns->cdev, &ns->cdev_device);
+
+	nvme_mpath_remove_sysfs_link(ns);
+
 	del_gendisk(ns->disk);
 
 	mutex_lock(&ns->ctrl->namespaces_lock);

drivers/nvme/host/multipath.c

@@ -686,6 +686,8 @@ static void nvme_mpath_set_live(struct nvme_ns *ns)
 		kblockd_schedule_work(&head->partition_scan_work);
 	}
 
+	nvme_mpath_add_sysfs_link(ns->head);
+
 	mutex_lock(&head->lock);
 	if (nvme_path_is_optimized(ns)) {
 		int node, srcu_idx;
@@ -768,6 +770,25 @@ static void nvme_update_ns_ana_state(struct nvme_ana_group_desc *desc,
 	if (nvme_state_is_live(ns->ana_state) &&
 	    nvme_ctrl_state(ns->ctrl) == NVME_CTRL_LIVE)
 		nvme_mpath_set_live(ns);
+	else {
+		/*
+		 * Add sysfs link from multipath head gendisk node to path
+		 * device gendisk node.
+		 * If path's ana state is live (i.e. state is either optimized
+		 * or non-optimized) while we alloc the ns then sysfs link would
+		 * be created from nvme_mpath_set_live(). In that case we would
+		 * not fallthrough this code path. However for the path's ana
+		 * state other than live, we call nvme_mpath_set_live() only
+		 * after ana state transitioned to the live state. But we still
+		 * want to create the sysfs link from head node to a path device
+		 * irrespctive of the path's ana state.
+		 * If we reach through here then it means that path's ana state
+		 * is not live but still create the sysfs link to this path from
+		 * head node if head node of the path has already come alive.
+		 */
+		if (test_bit(NVME_NSHEAD_DISK_LIVE, &ns->head->flags))
+			nvme_mpath_add_sysfs_link(ns->head);
+	}
 }
 
 static int nvme_update_ana_state(struct nvme_ctrl *ctrl,
@@ -967,6 +988,84 @@ static int nvme_lookup_ana_group_desc(struct nvme_ctrl *ctrl,
 	return -ENXIO; /* just break out of the loop */
 }
 
+void nvme_mpath_add_sysfs_link(struct nvme_ns_head *head)
+{
+	struct device *target;
+	int rc, srcu_idx;
+	struct nvme_ns *ns;
+	struct kobject *kobj;
+
+	/*
+	 * Ensure head disk node is already added otherwise we may get invalid
+	 * kobj for head disk node
+	 */
+	if (!test_bit(GD_ADDED, &head->disk->state))
+		return;
+
+	kobj = &disk_to_dev(head->disk)->kobj;
+
+	/*
+	 * loop through each ns chained through the head->list and create the
+	 * sysfs link from head node to the ns path node
+	 */
+	srcu_idx = srcu_read_lock(&head->srcu);
+
+	list_for_each_entry_rcu(ns, &head->list, siblings) {
+		/*
+		 * Avoid creating link if it already exists for the given path.
+		 * When path ana state transitions from optimized to non-
+		 * optimized or vice-versa, the nvme_mpath_set_live() is
+		 * invoked which in truns call this function. Now if the sysfs
+		 * link already exists for the given path and we attempt to re-
+		 * create the link then sysfs code would warn about it loudly.
+		 * So we evaluate NVME_NS_SYSFS_ATTR_LINK flag here to ensure
+		 * that we're not creating duplicate link.
+		 * The test_and_set_bit() is used because it is protecting
+		 * against multiple nvme paths being simultaneously added.
+		 */
+		if (test_and_set_bit(NVME_NS_SYSFS_ATTR_LINK, &ns->flags))
+			continue;
+
+		/*
+		 * Ensure that ns path disk node is already added otherwise we
+		 * may get invalid kobj name for target
+		 */
+		if (!test_bit(GD_ADDED, &ns->disk->state))
+			continue;
+
+		target = disk_to_dev(ns->disk);
+		/*
+		 * Create sysfs link from head gendisk kobject @kobj to the
+		 * ns path gendisk kobject @target->kobj.
+		 */
+		rc = sysfs_add_link_to_group(kobj, nvme_ns_mpath_attr_group.name,
+				&target->kobj, dev_name(target));
+		if (unlikely(rc)) {
+			dev_err(disk_to_dev(ns->head->disk),
+					"failed to create link to %s\n",
+					dev_name(target));
+			clear_bit(NVME_NS_SYSFS_ATTR_LINK, &ns->flags);
+		}
+	}
+
+	srcu_read_unlock(&head->srcu, srcu_idx);
+}
+
+void nvme_mpath_remove_sysfs_link(struct nvme_ns *ns)
+{
+	struct device *target;
+	struct kobject *kobj;
+
+	if (!test_bit(NVME_NS_SYSFS_ATTR_LINK, &ns->flags))
+		return;
+
+	target = disk_to_dev(ns->disk);
+	kobj = &disk_to_dev(ns->head->disk)->kobj;
+	sysfs_remove_link_from_group(kobj, nvme_ns_mpath_attr_group.name,
+			dev_name(target));
+	clear_bit(NVME_NS_SYSFS_ATTR_LINK, &ns->flags);
+}
+
 void nvme_mpath_add_disk(struct nvme_ns *ns, __le32 anagrpid)
 {
 	if (nvme_ctrl_use_ana(ns->ctrl)) {

drivers/nvme/host/nvme.h

@@ -534,10 +534,11 @@ struct nvme_ns {
 	struct nvme_ns_head *head;
 
 	unsigned long flags;
-#define NVME_NS_REMOVING	0
-#define NVME_NS_ANA_PENDING	2
-#define NVME_NS_FORCE_RO	3
-#define NVME_NS_READY		4
+#define NVME_NS_REMOVING		0
+#define NVME_NS_ANA_PENDING		2
+#define NVME_NS_FORCE_RO		3
+#define NVME_NS_READY			4
+#define NVME_NS_SYSFS_ATTR_LINK	5
 
 	struct cdev		cdev;
 	struct device		cdev_device;
@@ -933,6 +934,7 @@ int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo);
 int nvme_dev_uring_cmd(struct io_uring_cmd *ioucmd, unsigned int issue_flags);
 
 extern const struct attribute_group *nvme_ns_attr_groups[];
+extern const struct attribute_group nvme_ns_mpath_attr_group;
 extern const struct pr_ops nvme_pr_ops;
 extern const struct block_device_operations nvme_ns_head_ops;
 extern const struct attribute_group nvme_dev_attrs_group;
@@ -955,6 +957,8 @@ void nvme_mpath_default_iopolicy(struct nvme_subsystem *subsys);
 void nvme_failover_req(struct request *req);
 void nvme_kick_requeue_lists(struct nvme_ctrl *ctrl);
 int nvme_mpath_alloc_disk(struct nvme_ctrl *ctrl,struct nvme_ns_head *head);
+void nvme_mpath_add_sysfs_link(struct nvme_ns_head *ns);
+void nvme_mpath_remove_sysfs_link(struct nvme_ns *ns);
 void nvme_mpath_add_disk(struct nvme_ns *ns, __le32 anagrpid);
 void nvme_mpath_remove_disk(struct nvme_ns_head *head);
 int nvme_mpath_init_identify(struct nvme_ctrl *ctrl, struct nvme_id_ctrl *id);
@@ -1009,6 +1013,12 @@ static inline void nvme_mpath_add_disk(struct nvme_ns *ns, __le32 anagrpid)
 static inline void nvme_mpath_remove_disk(struct nvme_ns_head *head)
 {
 }
+static inline void nvme_mpath_add_sysfs_link(struct nvme_ns *ns)
+{
+}
+static inline void nvme_mpath_remove_sysfs_link(struct nvme_ns *ns)
+{
+}
 static inline bool nvme_mpath_clear_current_path(struct nvme_ns *ns)
 {
 	return false;

drivers/nvme/host/sysfs.c

@@ -299,8 +299,22 @@ static const struct attribute_group nvme_ns_attr_group = {
 	.is_visible	= nvme_ns_attrs_are_visible,
 };
 
+#ifdef CONFIG_NVME_MULTIPATH
+static struct attribute *nvme_ns_mpath_attrs[] = {
+	NULL,
+};
+
+const struct attribute_group nvme_ns_mpath_attr_group = {
+	.name           = "multipath",
+	.attrs		= nvme_ns_mpath_attrs,
+};
+#endif
+
 const struct attribute_group *nvme_ns_attr_groups[] = {
 	&nvme_ns_attr_group,
+#ifdef CONFIG_NVME_MULTIPATH
+	&nvme_ns_mpath_attr_group,
+#endif
 	NULL,
 };