target: Convert se_node_acl->device_list[] to RCU hlist

This patch converts se_node_acl->device_list[] table for mappedluns
to modern RCU hlist_head usage in order to support an arbitrary number
of node_acl lun mappings.

It converts transport_lookup_*_lun() fast-path code to use RCU read path
primitives when looking up se_dev_entry.  It adds a new hlist_head at
se_node_acl->lun_entry_hlist for this purpose.

For transport_lookup_cmd_lun() code, it works with existing per-cpu
se_lun->lun_ref when associating se_cmd with se_lun + se_device.
Also, go ahead and update core_create_device_list_for_node() +
core_free_device_list_for_node() to use ->lun_entry_hlist.

It also converts se_dev_entry->pr_ref_count access to use modern
struct kref counting, and updates core_disable_device_list_for_node()
to kref_put() and block on se_deve->pr_comp waiting for outstanding PR
special-case PR references to drop, then invoke kfree_rcu() to wait
for the RCU grace period to complete before releasing memory.

So now that se_node_acl->lun_entry_hlist fast path access uses RCU
protected pointers, go ahead and convert remaining non-fast path
RCU updater code using ->lun_entry_lock to struct mutex to allow
callers to block while walking se_node_acl->lun_entry_hlist.

Finally drop the left-over core_clear_initiator_node_from_tpg() that
originally cleared lun_access during se_node_acl shutdown, as post
RCU conversion it now becomes duplicated logic.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Sagi Grimberg <sagig@mellanox.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>

drivers/target/target_core_alua.c

@@ -1001,7 +1001,8 @@ static void core_alua_do_transition_tg_pt_work(struct work_struct *work)
 		spin_lock_bh(&port->sep_alua_lock);
 		list_for_each_entry(se_deve, &port->sep_alua_list,
 					alua_port_list) {
-			lacl = se_deve->se_lun_acl;
+			lacl = rcu_dereference_check(se_deve->se_lun_acl,
+					     lockdep_is_held(&port->sep_alua_lock));
 			/*
 			 * se_deve->se_lun_acl pointer may be NULL for a
 			 * entry created without explicit Node+MappedLUN ACLs

drivers/target/target_core_fabric_configfs.c

@@ -123,16 +123,16 @@ static int target_fabric_mappedlun_link(
 	 * which be will write protected (READ-ONLY) when
 	 * tpg_1/attrib/demo_mode_write_protect=1
 	 */
-	spin_lock_irq(&lacl->se_lun_nacl->device_list_lock);
-	deve = lacl->se_lun_nacl->device_list[lacl->mapped_lun];
-	if (deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS)
+	rcu_read_lock();
+	deve = target_nacl_find_deve(lacl->se_lun_nacl, lacl->mapped_lun);
+	if (deve)
 		lun_access = deve->lun_flags;
 	else
 		lun_access =
 			(se_tpg->se_tpg_tfo->tpg_check_prod_mode_write_protect(
 				se_tpg)) ? TRANSPORT_LUNFLAGS_READ_ONLY :
 					   TRANSPORT_LUNFLAGS_READ_WRITE;
-	spin_unlock_irq(&lacl->se_lun_nacl->device_list_lock);
+	rcu_read_unlock();
 	/*
 	 * Determine the actual mapped LUN value user wants..
 	 *
@@ -149,23 +149,13 @@ static int target_fabric_mappedlun_unlink(
 	struct config_item *lun_acl_ci,
 	struct config_item *lun_ci)
 {
-	struct se_lun *lun;
 	struct se_lun_acl *lacl = container_of(to_config_group(lun_acl_ci),
 			struct se_lun_acl, se_lun_group);
-	struct se_node_acl *nacl = lacl->se_lun_nacl;
-	struct se_dev_entry *deve = nacl->device_list[lacl->mapped_lun];
-	struct se_portal_group *se_tpg;
-	/*
-	 * Determine if the underlying MappedLUN has already been released..
-	 */
-	if (!deve->se_lun)
-		return 0;
-
-	lun = container_of(to_config_group(lun_ci), struct se_lun, lun_group);
-	se_tpg = lun->lun_sep->sep_tpg;
+	struct se_lun *lun = container_of(to_config_group(lun_ci),
+			struct se_lun, lun_group);
+	struct se_portal_group *se_tpg = lun->lun_sep->sep_tpg;
 
-	core_dev_del_initiator_node_lun_acl(se_tpg, lun, lacl);
-	return 0;
+	return core_dev_del_initiator_node_lun_acl(se_tpg, lun, lacl);
 }
 
 CONFIGFS_EATTR_STRUCT(target_fabric_mappedlun, se_lun_acl);
@@ -181,14 +171,15 @@ static ssize_t target_fabric_mappedlun_show_write_protect(
 {
 	struct se_node_acl *se_nacl = lacl->se_lun_nacl;
 	struct se_dev_entry *deve;
-	ssize_t len;
+	ssize_t len = 0;
 
-	spin_lock_irq(&se_nacl->device_list_lock);
-	deve = se_nacl->device_list[lacl->mapped_lun];
+	rcu_read_lock();
+	deve = target_nacl_find_deve(se_nacl, lacl->mapped_lun);
+	if (deve) {
 		len = sprintf(page, "%d\n",
-			(deve->lun_flags & TRANSPORT_LUNFLAGS_READ_ONLY) ?
-			1 : 0);
-	spin_unlock_irq(&se_nacl->device_list_lock);
+			(deve->lun_flags & TRANSPORT_LUNFLAGS_READ_ONLY) ? 1 : 0);
+	}
+	rcu_read_unlock();
 
 	return len;
 }

drivers/target/target_core_internal.h

@@ -9,13 +9,15 @@ extern struct mutex g_device_mutex;
 extern struct list_head g_device_list;
 
 struct se_dev_entry *core_get_se_deve_from_rtpi(struct se_node_acl *, u16);
-int	core_free_device_list_for_node(struct se_node_acl *,
+void	target_pr_kref_release(struct kref *);
+void	core_free_device_list_for_node(struct se_node_acl *,
 		struct se_portal_group *);
 void	core_update_device_list_access(u32, u32, struct se_node_acl *);
+struct se_dev_entry *target_nacl_find_deve(struct se_node_acl *, u32);
 int	core_enable_device_list_for_node(struct se_lun *, struct se_lun_acl *,
 		u32, u32, struct se_node_acl *, struct se_portal_group *);
-int	core_disable_device_list_for_node(struct se_lun *, struct se_lun_acl *,
-		u32, u32, struct se_node_acl *, struct se_portal_group *);
+void	core_disable_device_list_for_node(struct se_lun *, struct se_dev_entry *,
+		struct se_node_acl *, struct se_portal_group *);
 void	core_clear_lun_from_tpg(struct se_lun *, struct se_portal_group *);
 int	core_dev_export(struct se_device *, struct se_portal_group *,
 		struct se_lun *);

drivers/target/target_core_pscsi.c

@@ -47,6 +47,7 @@
 #include <target/target_core_backend_configfs.h>
 
 #include "target_core_alua.h"
+#include "target_core_internal.h"
 #include "target_core_pscsi.h"
 
 #define ISPRINT(a)  ((a >= ' ') && (a <= '~'))
@@ -637,12 +638,14 @@ static void pscsi_transport_complete(struct se_cmd *cmd, struct scatterlist *sg,
 	 * Hack to make sure that Write-Protect modepage is set if R/O mode is
 	 * forced.
 	 */
-	if (!cmd->se_deve || !cmd->data_length)
+	if (!cmd->data_length)
 		goto after_mode_sense;
 
 	if (((cdb[0] == MODE_SENSE) || (cdb[0] == MODE_SENSE_10)) &&
 	     (status_byte(result) << 1) == SAM_STAT_GOOD) {
-		if (cmd->se_deve->lun_flags & TRANSPORT_LUNFLAGS_READ_ONLY) {
+		bool read_only = target_lun_is_rdonly(cmd);
+
+		if (read_only) {
 			unsigned char *buf;
 
 			buf = transport_kmap_data_sg(cmd);

drivers/target/target_core_spc.c

@@ -981,6 +981,7 @@ static sense_reason_t spc_emulate_modesense(struct se_cmd *cmd)
 	int length = 0;
 	int ret;
 	int i;
+	bool read_only = target_lun_is_rdonly(cmd);;
 
 	memset(buf, 0, SE_MODE_PAGE_BUF);
 
@@ -991,9 +992,7 @@ static sense_reason_t spc_emulate_modesense(struct se_cmd *cmd)
 	length = ten ? 3 : 2;
 
 	/* DEVICE-SPECIFIC PARAMETER */
-	if ((cmd->se_lun->lun_access & TRANSPORT_LUNFLAGS_READ_ONLY) ||
-	    (cmd->se_deve &&
-	     (cmd->se_deve->lun_flags & TRANSPORT_LUNFLAGS_READ_ONLY)))
+	if ((cmd->se_lun->lun_access & TRANSPORT_LUNFLAGS_READ_ONLY) || read_only)
 		spc_modesense_write_protect(&buf[length], type);
 
 	/*
@@ -1211,8 +1210,9 @@ sense_reason_t spc_emulate_report_luns(struct se_cmd *cmd)
 {
 	struct se_dev_entry *deve;
 	struct se_session *sess = cmd->se_sess;
+	struct se_node_acl *nacl;
 	unsigned char *buf;
-	u32 lun_count = 0, offset = 8, i;
+	u32 lun_count = 0, offset = 8;
 
 	if (cmd->data_length < 16) {
 		pr_warn("REPORT LUNS allocation length %u too small\n",
@@ -1234,12 +1234,10 @@ sense_reason_t spc_emulate_report_luns(struct se_cmd *cmd)
 		lun_count = 1;
 		goto done;
 	}
+	nacl = sess->se_node_acl;
 
-	spin_lock_irq(&sess->se_node_acl->device_list_lock);
-	for (i = 0; i < TRANSPORT_MAX_LUNS_PER_TPG; i++) {
-		deve = sess->se_node_acl->device_list[i];
-		if (!(deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS))
-			continue;
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(deve, &nacl->lun_entry_hlist, link) {
 		/*
 		 * We determine the correct LUN LIST LENGTH even once we
 		 * have reached the initial allocation length.
@@ -1252,7 +1250,7 @@ sense_reason_t spc_emulate_report_luns(struct se_cmd *cmd)
 		int_to_scsilun(deve->mapped_lun, (struct scsi_lun *)&buf[offset]);
 		offset += 8;
 	}
-	spin_unlock_irq(&sess->se_node_acl->device_list_lock);
+	rcu_read_unlock();
 
 	/*
 	 * See SPC3 r07, page 159.

drivers/target/target_core_tpg.c

@@ -47,42 +47,6 @@ extern struct se_device *g_lun0_dev;
 static DEFINE_SPINLOCK(tpg_lock);
 static LIST_HEAD(tpg_list);
 
-/*	core_clear_initiator_node_from_tpg():
- *
- *
- */
-static void core_clear_initiator_node_from_tpg(
-	struct se_node_acl *nacl,
-	struct se_portal_group *tpg)
-{
-	int i;
-	struct se_dev_entry *deve;
-	struct se_lun *lun;
-
-	spin_lock_irq(&nacl->device_list_lock);
-	for (i = 0; i < TRANSPORT_MAX_LUNS_PER_TPG; i++) {
-		deve = nacl->device_list[i];
-
-		if (!(deve->lun_flags & TRANSPORT_LUNFLAGS_INITIATOR_ACCESS))
-			continue;
-
-		if (!deve->se_lun) {
-			pr_err("%s device entries device pointer is"
-				" NULL, but Initiator has access.\n",
-				tpg->se_tpg_tfo->get_fabric_name());
-			continue;
-		}
-
-		lun = deve->se_lun;
-		spin_unlock_irq(&nacl->device_list_lock);
-		core_disable_device_list_for_node(lun, NULL, deve->mapped_lun,
-			TRANSPORT_LUNFLAGS_NO_ACCESS, nacl, tpg);
-
-		spin_lock_irq(&nacl->device_list_lock);
-	}
-	spin_unlock_irq(&nacl->device_list_lock);
-}
-
 /*	__core_tpg_get_initiator_node_acl():
  *
  *	spin_lock_bh(&tpg->acl_node_lock); must be held when calling
@@ -225,35 +189,6 @@ static void *array_zalloc(int n, size_t size, gfp_t flags)
 	return a;
 }
 
-/*      core_create_device_list_for_node():
- *
- *
- */
-static int core_create_device_list_for_node(struct se_node_acl *nacl)
-{
-	struct se_dev_entry *deve;
-	int i;
-
-	nacl->device_list = array_zalloc(TRANSPORT_MAX_LUNS_PER_TPG,
-			sizeof(struct se_dev_entry), GFP_KERNEL);
-	if (!nacl->device_list) {
-		pr_err("Unable to allocate memory for"
-			" struct se_node_acl->device_list\n");
-		return -ENOMEM;
-	}
-	for (i = 0; i < TRANSPORT_MAX_LUNS_PER_TPG; i++) {
-		deve = nacl->device_list[i];
-
-		atomic_set(&deve->ua_count, 0);
-		atomic_set(&deve->pr_ref_count, 0);
-		spin_lock_init(&deve->ua_lock);
-		INIT_LIST_HEAD(&deve->alua_port_list);
-		INIT_LIST_HEAD(&deve->ua_list);
-	}
-
-	return 0;
-}
-
 static struct se_node_acl *target_alloc_node_acl(struct se_portal_group *tpg,
 		const unsigned char *initiatorname)
 {
@@ -266,10 +201,11 @@ static struct se_node_acl *target_alloc_node_acl(struct se_portal_group *tpg,
 
 	INIT_LIST_HEAD(&acl->acl_list);
 	INIT_LIST_HEAD(&acl->acl_sess_list);
+	INIT_HLIST_HEAD(&acl->lun_entry_hlist);
 	kref_init(&acl->acl_kref);
 	init_completion(&acl->acl_free_comp);
-	spin_lock_init(&acl->device_list_lock);
 	spin_lock_init(&acl->nacl_sess_lock);
+	mutex_init(&acl->lun_entry_mutex);
 	atomic_set(&acl->acl_pr_ref_count, 0);
 	if (tpg->se_tpg_tfo->tpg_get_default_depth)
 		acl->queue_depth = tpg->se_tpg_tfo->tpg_get_default_depth(tpg);
@@ -281,15 +217,11 @@ static struct se_node_acl *target_alloc_node_acl(struct se_portal_group *tpg,
 
 	tpg->se_tpg_tfo->set_default_node_attributes(acl);
 
-	if (core_create_device_list_for_node(acl) < 0)
-		goto out_free_acl;
 	if (core_set_queue_depth_for_node(tpg, acl) < 0)
-		goto out_free_device_list;
+		goto out_free_acl;
 
 	return acl;
 
-out_free_device_list:
-	core_free_device_list_for_node(acl, tpg);
 out_free_acl:
 	kfree(acl);
 	return NULL;
@@ -454,7 +386,6 @@ void core_tpg_del_initiator_node_acl(struct se_node_acl *acl)
 	wait_for_completion(&acl->acl_free_comp);
 
 	core_tpg_wait_for_nacl_pr_ref(acl);
-	core_clear_initiator_node_from_tpg(acl, tpg);
 	core_free_device_list_for_node(acl, tpg);
 
 	pr_debug("%s_TPG[%hu] - Deleted ACL with TCQ Depth: %d for %s"

drivers/target/target_core_ua.c

@@ -50,9 +50,17 @@ target_scsi3_ua_check(struct se_cmd *cmd)
 	if (!nacl)
 		return 0;
 
-	deve = nacl->device_list[cmd->orig_fe_lun];
-	if (!atomic_read(&deve->ua_count))
+	rcu_read_lock();
+	deve = target_nacl_find_deve(nacl, cmd->orig_fe_lun);
+	if (!deve) {
+		rcu_read_unlock();
 		return 0;
+	}
+	if (!atomic_read(&deve->ua_count)) {
+		rcu_read_unlock();
+		return 0;
+	}
+	rcu_read_unlock();
 	/*
 	 * From sam4r14, section 5.14 Unit attention condition:
 	 *
@@ -103,9 +111,12 @@ int core_scsi3_ua_allocate(
 	ua->ua_asc = asc;
 	ua->ua_ascq = ascq;
 
-	spin_lock_irq(&nacl->device_list_lock);
-	deve = nacl->device_list[unpacked_lun];
-
+	rcu_read_lock();
+	deve = target_nacl_find_deve(nacl, unpacked_lun);
+	if (!deve) {
+		rcu_read_unlock();
+		return -EINVAL;
+	}
 	spin_lock(&deve->ua_lock);
 	list_for_each_entry_safe(ua_p, ua_tmp, &deve->ua_list, ua_nacl_list) {
 		/*
@@ -113,7 +124,7 @@ int core_scsi3_ua_allocate(
 		 */
 		if ((ua_p->ua_asc == asc) && (ua_p->ua_ascq == ascq)) {
 			spin_unlock(&deve->ua_lock);
-			spin_unlock_irq(&nacl->device_list_lock);
+			rcu_read_unlock();
 			kmem_cache_free(se_ua_cache, ua);
 			return 0;
 		}
@@ -158,14 +169,13 @@ int core_scsi3_ua_allocate(
 			list_add_tail(&ua->ua_nacl_list,
 				&deve->ua_list);
 		spin_unlock(&deve->ua_lock);
-		spin_unlock_irq(&nacl->device_list_lock);
 
 		atomic_inc_mb(&deve->ua_count);
+		rcu_read_unlock();
 		return 0;
 	}
 	list_add_tail(&ua->ua_nacl_list, &deve->ua_list);
 	spin_unlock(&deve->ua_lock);
-	spin_unlock_irq(&nacl->device_list_lock);
 
 	pr_debug("[%s]: Allocated UNIT ATTENTION, mapped LUN: %u, ASC:"
 		" 0x%02x, ASCQ: 0x%02x\n",
@@ -173,6 +183,7 @@ int core_scsi3_ua_allocate(
 		asc, ascq);
 
 	atomic_inc_mb(&deve->ua_count);
+	rcu_read_unlock();
 	return 0;
 }
 
@@ -210,10 +221,14 @@ void core_scsi3_ua_for_check_condition(
 	if (!nacl)
 		return;
 
-	spin_lock_irq(&nacl->device_list_lock);
-	deve = nacl->device_list[cmd->orig_fe_lun];
+	rcu_read_lock();
+	deve = target_nacl_find_deve(nacl, cmd->orig_fe_lun);
+	if (!deve) {
+		rcu_read_unlock();
+		return;
+	}
 	if (!atomic_read(&deve->ua_count)) {
-		spin_unlock_irq(&nacl->device_list_lock);
+		rcu_read_unlock();
 		return;
 	}
 	/*
@@ -249,7 +264,7 @@ void core_scsi3_ua_for_check_condition(
 		atomic_dec_mb(&deve->ua_count);
 	}
 	spin_unlock(&deve->ua_lock);
-	spin_unlock_irq(&nacl->device_list_lock);
+	rcu_read_unlock();
 
 	pr_debug("[%s]: %s UNIT ATTENTION condition with"
 		" INTLCK_CTRL: %d, mapped LUN: %u, got CDB: 0x%02x"
@@ -278,10 +293,14 @@ int core_scsi3_ua_clear_for_request_sense(
 	if (!nacl)
 		return -EINVAL;
 
-	spin_lock_irq(&nacl->device_list_lock);
-	deve = nacl->device_list[cmd->orig_fe_lun];
+	rcu_read_lock();
+	deve = target_nacl_find_deve(nacl, cmd->orig_fe_lun);
+	if (!deve) {
+		rcu_read_unlock();
+		return -EINVAL;
+	}
 	if (!atomic_read(&deve->ua_count)) {
-		spin_unlock_irq(&nacl->device_list_lock);
+		rcu_read_unlock();
 		return -EPERM;
 	}
 	/*
@@ -307,7 +326,7 @@ int core_scsi3_ua_clear_for_request_sense(
 		atomic_dec_mb(&deve->ua_count);
 	}
 	spin_unlock(&deve->ua_lock);
-	spin_unlock_irq(&nacl->device_list_lock);
+	rcu_read_unlock();
 
 	pr_debug("[%s]: Released UNIT ATTENTION condition, mapped"
 		" LUN: %u, got REQUEST_SENSE reported ASC: 0x%02x,"

include/target/target_core_backend.h

@@ -101,7 +101,7 @@ int	target_alloc_sgl(struct scatterlist **, unsigned int *, u32, bool);
 sense_reason_t	transport_generic_map_mem_to_cmd(struct se_cmd *,
 		struct scatterlist *, u32, struct scatterlist *, u32);
 
-void	array_free(void *array, int n);
+bool	target_lun_is_rdonly(struct se_cmd *);
 
 /* From target_core_configfs.c to setup default backend config_item_types */
 void	target_core_setup_sub_cits(struct se_subsystem_api *);

include/target/target_core_base.h

@@ -160,10 +160,8 @@ enum se_cmd_flags_table {
 
 /* struct se_dev_entry->lun_flags and struct se_lun->lun_access */
 enum transport_lunflags_table {
-	TRANSPORT_LUNFLAGS_NO_ACCESS		= 0x00,
-	TRANSPORT_LUNFLAGS_INITIATOR_ACCESS	= 0x01,
-	TRANSPORT_LUNFLAGS_READ_ONLY		= 0x02,
-	TRANSPORT_LUNFLAGS_READ_WRITE		= 0x04,
+	TRANSPORT_LUNFLAGS_READ_ONLY		= 0x01,
+	TRANSPORT_LUNFLAGS_READ_WRITE		= 0x02,
 };
 
 /*
@@ -584,10 +582,10 @@ struct se_node_acl {
 	char			acl_tag[MAX_ACL_TAG_SIZE];
 	/* Used for PR SPEC_I_PT=1 and REGISTER_AND_MOVE */
 	atomic_t		acl_pr_ref_count;
-	struct se_dev_entry	**device_list;
+	struct hlist_head	lun_entry_hlist;
 	struct se_session	*nacl_sess;
 	struct se_portal_group *se_tpg;
-	spinlock_t		device_list_lock;
+	struct mutex		lun_entry_mutex;
 	spinlock_t		nacl_sess_lock;
 	struct config_group	acl_group;
 	struct config_group	acl_attrib_group;
@@ -644,20 +642,23 @@ struct se_dev_entry {
 	/* See transport_lunflags_table */
 	u32			lun_flags;
 	u32			mapped_lun;
-	u32			total_cmds;
 	u64			pr_res_key;
 	u64			creation_time;
 	u32			attach_count;
-	u64			read_bytes;
-	u64			write_bytes;
+	atomic_long_t		total_cmds;
+	atomic_long_t		read_bytes;
+	atomic_long_t		write_bytes;
 	atomic_t		ua_count;
 	/* Used for PR SPEC_I_PT=1 and REGISTER_AND_MOVE */
-	atomic_t		pr_ref_count;
-	struct se_lun_acl	*se_lun_acl;
+	struct kref		pr_kref;
+	struct completion	pr_comp;
+	struct se_lun_acl __rcu	*se_lun_acl;
 	spinlock_t		ua_lock;
-	struct se_lun		*se_lun;
+	struct se_lun __rcu	*se_lun;
 	struct list_head	alua_port_list;
 	struct list_head	ua_list;
+	struct hlist_node	link;
+	struct rcu_head		rcu_head;
 };
 
 struct se_dev_attrib {
@@ -703,6 +704,7 @@ struct se_port_stat_grps {
 };
 
 struct se_lun {
+	u16			lun_rtpi;
 #define SE_LUN_LINK_MAGIC			0xffff7771
 	u32			lun_link_magic;
 	/* See transport_lun_status_table */
@@ -710,6 +712,7 @@ struct se_lun {
 	u32			lun_access;
 	u32			lun_flags;
 	u32			unpacked_lun;
+	u32			lun_index;
 	atomic_t		lun_acl_count;
 	spinlock_t		lun_acl_lock;
 	spinlock_t		lun_sep_lock;