[PATCH v4 1/2] nvme: add per-queue sysfs directories

[Saravanan D <saravanand@crusoe.ai>]

Transports have per I/O queue state worth exposing, the nvme-tcp
io_cpu to begin with, but the core has no representation of an
individual queue and no place under the controller device to hang
per queue attributes.

Add a small nvme_queue_info that a transport embeds in its queue
structure, and helpers that register it as

  /sys/class/nvme/nvmeX/queues/<qid>/

with a transport provided ktype. The queues directory is created on
the first registration and released with the controller device.
Registration is expected once per controller lifetime, so the
directories persist while the transport's queues cycle across
reconnects.

Suggested-by: Sagi Grimberg <sagi@grimberg.me>
Link: https://lore.kernel.org/linux-nvme/13c160d0-0657-47ce-9a16-58d348267ab0@grimberg.me/
Assisted-by: Claude:claude-opus-4-8 [Claude Code]
Signed-off-by: Saravanan D <saravanand@crusoe.ai>
---
 drivers/nvme/host/core.c  |  2 ++
 drivers/nvme/host/nvme.h  | 23 +++++++++++++++++++++
 drivers/nvme/host/sysfs.c | 42 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 67 insertions(+)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index beea23d04a70..211d05a1cc48 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -5197,6 +5197,8 @@ static void nvme_free_ctrl(struct device *dev)
 
 	ctrl->ops->free_ctrl(ctrl);
 
+	kobject_put(ctrl->queues_kobj);
+
 	if (subsys)
 		nvme_put_subsystem(subsys);
 }
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index bac25b287d25..20410f0d6118 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -339,6 +339,28 @@ enum nvme_ctrl_flags {
 	NVME_CTRL_FROZEN		= 6,
 };
 
+enum nvme_queue_info_flags {
+	NVME_QUEUE_INFO_REGISTERED	= 0,
+	NVME_QUEUE_INFO_IO_CPU_USER	= 1,
+};
+
+/*
+ * Common per-queue state, embedded in the transport's queue structure and
+ * backing the /sys/class/nvme/nvmeX/queues/<qid>/ directory.
+ */
+struct nvme_queue_info {
+	struct kobject		kobj;
+	unsigned int		qid;
+	unsigned long		flags;
+};
+
+struct nvme_ctrl;
+
+int nvme_register_queue_info(struct nvme_ctrl *ctrl,
+			     struct nvme_queue_info *qinfo, unsigned int qid,
+			     const struct kobj_type *ktype);
+void nvme_unregister_queue_info(struct nvme_queue_info *qinfo);
+
 struct nvme_ctrl {
 	bool comp_seen;
 	bool identified;
@@ -360,6 +382,7 @@ struct nvme_ctrl {
 	struct srcu_struct srcu;
 	struct device ctrl_device;
 	struct device *device;	/* char device */
+	struct kobject *queues_kobj;
 #ifdef CONFIG_NVME_HWMON
 	struct device *hwmon_device;
 #endif
diff --git a/drivers/nvme/host/sysfs.c b/drivers/nvme/host/sysfs.c
index 4ac3ea3a86fd..a2780445c33e 100644
--- a/drivers/nvme/host/sysfs.c
+++ b/drivers/nvme/host/sysfs.c
@@ -1313,3 +1313,45 @@ const struct attribute_group *nvme_subsys_attrs_groups[] = {
 	&nvme_subsys_attrs_group,
 	NULL,
 };
+
+/*
+ * Per-queue sysfs directories under /sys/class/nvme/nvmeX/queues/. The
+ * transport backing the controller registers a directory for each of its
+ * queues and provides the attributes through the ktype. Registration is
+ * expected once per controller lifetime, so the directories persist while
+ * the transport's queues cycle across reconnects. The queues directory
+ * itself is released with the controller device.
+ */
+int nvme_register_queue_info(struct nvme_ctrl *ctrl,
+			     struct nvme_queue_info *qinfo, unsigned int qid,
+			     const struct kobj_type *ktype)
+{
+	int ret;
+
+	if (!ctrl->queues_kobj) {
+		ctrl->queues_kobj = kobject_create_and_add("queues",
+							   &ctrl->device->kobj);
+		if (!ctrl->queues_kobj)
+			return -ENOMEM;
+	}
+
+	qinfo->qid = qid;
+	ret = kobject_init_and_add(&qinfo->kobj, ktype, ctrl->queues_kobj,
+				   "%u", qid);
+	if (ret) {
+		kobject_put(&qinfo->kobj);
+		return ret;
+	}
+
+	set_bit(NVME_QUEUE_INFO_REGISTERED, &qinfo->flags);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(nvme_register_queue_info);
+
+void nvme_unregister_queue_info(struct nvme_queue_info *qinfo)
+{
+	if (!test_and_clear_bit(NVME_QUEUE_INFO_REGISTERED, &qinfo->flags))
+		return;
+	kobject_put(&qinfo->kobj);
+}
+EXPORT_SYMBOL_GPL(nvme_unregister_queue_info);
-- 
2.55.0