nvmet: implement unique discovery NQN
Subsystems (2)
PATCH v2 1/1 nvmet: implement unique discovery NQN
From: Hannes Reinecke <[email protected]>
Add a configfs attribute 'discovery_nqn' in the 'nvmet' configfs directory to specify the unique discovery NQN.
--- drivers/nvme/target/configfs.c | 46 ++++++++++++++++++++++++++++++++++ drivers/nvme/target/core.c | 7 ++++++ 2 files changed, 53 insertions(+)
drivers/nvme/target/configfs.c +46 -0
@@ -1587,6 +1587,11 @@ static struct config_group *nvmet_subsys_make(struct config_group *group,return ERR_PTR (-EINVAL ); } +if (sysfs_streq (name, nvmet_disc_subsys->subsysnqn)) { +pr_err ("can't create subsystem using unique discovery NQN\n" ); +return ERR_PTR (-EINVAL ); + } + subsys =nvmet_subsys_alloc (name,NVME_NQN_NVME );if (IS_ERR (subsys))return ERR_CAST (subsys);
@@ -2131,7 +2136,48 @@ static const struct config_item_type nvmet_hosts_type = {static struct config_group nvmet_hosts_group; +static ssize_t nvmet_root_discovery_nqn_show (struct config_item *item, +char *page) +{ +return sprintf (page,"%s\n" , nvmet_disc_subsys->subsysnqn); +} + +static ssize_t nvmet_root_discovery_nqn_store (struct config_item *item, +const char *page,size_t count) +{ +struct list_head *entry; +size_t len; + + len =strcspn (page,"\n" ); +if (!len || len >NVMF_NQN_FIELD_LEN -1 ) +return -EINVAL ; + +down_write (&nvmet_config_sem); +list_for_each (entry, &nvmet_subsystems_group.cg_children) { +struct config_item *item = + container_of(entry, struct config_item, ci_entry); +if (!strncmp (config_item_name (item), page, len)) { +pr_err ("duplicate NQN %s\n" ,config_item_name (item)); +up_write (&nvmet_config_sem); +return -EINVAL ; + } + } +memset (nvmet_disc_subsys->subsysnqn,0 ,NVMF_NQN_FIELD_LEN ); +memcpy (nvmet_disc_subsys->subsysnqn, page, len); +up_write (&nvmet_config_sem); + +return len; +} + +CONFIGFS_ATTR (nvmet_root_, discovery_nqn); + +static struct configfs_attribute *nvmet_root_attrs[] = { + &nvmet_root_attr_discovery_nqn, +NULL , +}; +static const struct config_item_type nvmet_root_type = { + .ct_attrs = nvmet_root_attrs, .ct_owner =THIS_MODULE , };
drivers/nvme/target/core.c +7 -0
@@ -1530,6 +1530,13 @@ static struct nvmet_subsys *nvmet_find_get_subsys(struct nvmet_port *port, }down_read (&nvmet_config_sem); +if (!strncmp (nvmet_disc_subsys->subsysnqn, subsysnqn, +NVMF_NQN_SIZE )) { +if (kref_get_unless_zero (&nvmet_disc_subsys->ref)) { +up_read (&nvmet_config_sem); +return nvmet_disc_subsys; + } + }list_for_each_entry (p, &port->subsystems, entry) {if (!strncmp (p->subsys->subsysnqn, subsysnqn,NVMF_NQN_SIZE )) {
Signature · 1 line
2.35.3
Christoph Hellwig
maintainerQuoted 4 lines · On Fri, Dec 15, 2023 at 01:39:29PM +0100, [email protected] wrote:
From: Hannes Reinecke <[email protected]>
Add a configfs attribute '
discovery_nqn' in the 'nvmet' configfs directory to specify the unique discovery NQN.
That's a bit of a sparse commit log. What do unique discovery NQN do? Please include a reference and a bit of an explnation. And tell us why this feature is useful for Linux to have.
Sagi Grimberg
maintainerQuoted 7 lines
Quoted 3 lines
Add a configfs attribute '
discovery_nqn' in the 'nvmet' configfs directory to specify the unique discovery NQN.That's a bit of a sparse commit log. What do unique discovery NQN do? Please include a reference and a bit of an explnation. And tell us why this feature is useful for Linux to have.
I think that a unique NQN is better to authenticate against. That would be the motivation in my mind here.
Quoted 10 lines · Sagi Grimberg
Quoted 7 lines
Quoted 3 lines
Add a configfs attribute '
discovery_nqn' in the 'nvmet' configfs directory to specify the unique discovery NQN.That's a bit of a sparse commit log. What do unique discovery NQN do? Please include a reference and a bit of an explnation. And tell us why this feature is useful for Linux to have.
I think that a unique NQN is better to authenticate against. That would be the motivation in my mind here.
Correct; unique NQN is actually a requirement for authentication. I guess there is even a TPAR out there to make it so.
But my main driver here is to test persistent discovery controllers as this impacts far more than 'just' the nvme subsystem. Most notably it requires enablement from the OS side, and possibly modifications in various areas (udev, systemd, dracut, etc).
All of which need to be tested.
Cheers,
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich