scsi: target: Add iscsi/cpus_allowed_list in configfs

The RX/TX threads for iSCSI connection can be scheduled to any online CPUs,
and will not be rescheduled.

When binding other heavy load threads along with iSCSI connection RX/TX
thread to the same CPU, the iSCSI performance will be worse.

Add iscsi/cpus_allowed_list in configfs. The available CPU set of iSCSI
connection RX/TX threads is allowed_cpus & online_cpus. If it is modified,
all RX/TX threads will be rescheduled.

Link: https://lore.kernel.org/r/20220301075500.14266-1-mingzhe.zou@easystack.cn
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Mingzhe Zou <mingzhe.zou@easystack.cn>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

3 files changed, 114 insertions, 3 deletions

diff --git a/drivers/target/iscsi/iscsi_target.c b/drivers/target/iscsi/iscsi_target.c
index 2c54c5d8412d8..6fe6a6bab3f46 100644
--- a/drivers/target/iscsi/iscsi_target.c
+++ b/drivers/target/iscsi/iscsi_target.c
@@ -702,13 +702,19 @@ static int __init iscsi_target_init_module(void)
 	if (!iscsit_global->ts_bitmap)
 		goto configfs_out;
 
+	if (!zalloc_cpumask_var(&iscsit_global->allowed_cpumask, GFP_KERNEL)) {
+		pr_err("Unable to allocate iscsit_global->allowed_cpumask\n");
+		goto bitmap_out;
+	}
+	cpumask_setall(iscsit_global->allowed_cpumask);
+
 	lio_qr_cache = kmem_cache_create("lio_qr_cache",
 			sizeof(struct iscsi_queue_req),
 			__alignof__(struct iscsi_queue_req), 0, NULL);
 	if (!lio_qr_cache) {
 		pr_err("Unable to kmem_cache_create() for"
 				" lio_qr_cache\n");
-		goto bitmap_out;
+		goto cpumask_out;
 	}
 
 	lio_dr_cache = kmem_cache_create("lio_dr_cache",
@@ -753,6 +759,8 @@ dr_out:
 	kmem_cache_destroy(lio_dr_cache);
 qr_out:
 	kmem_cache_destroy(lio_qr_cache);
+cpumask_out:
+	free_cpumask_var(iscsit_global->allowed_cpumask);
 bitmap_out:
 	vfree(iscsit_global->ts_bitmap);
 configfs_out:
@@ -782,6 +790,7 @@ static void __exit iscsi_target_cleanup_module(void)
 
 	target_unregister_template(&iscsi_ops);
 
+	free_cpumask_var(iscsit_global->allowed_cpumask);
 	vfree(iscsit_global->ts_bitmap);
 	kfree(iscsit_global);
 }
@@ -3587,6 +3596,11 @@ static int iscsit_send_reject(
 void iscsit_thread_get_cpumask(struct iscsi_conn *conn)
 {
 	int ord, cpu;
+	cpumask_t conn_allowed_cpumask;
+
+	cpumask_and(&conn_allowed_cpumask, iscsit_global->allowed_cpumask,
+		    cpu_online_mask);
+
 	/*
 	 * bitmap_id is assigned from iscsit_global->ts_bitmap from
 	 * within iscsit_start_kthreads()
@@ -3595,8 +3609,9 @@ void iscsit_thread_get_cpumask(struct iscsi_conn *conn)
 	 * iSCSI connection's RX/TX threads will be scheduled to
 	 * execute upon.
 	 */
-	ord = conn->bitmap_id % cpumask_weight(cpu_online_mask);
-	for_each_online_cpu(cpu) {
+	cpumask_clear(conn->conn_cpumask);
+	ord = conn->bitmap_id % cpumask_weight(&conn_allowed_cpumask);
+	for_each_cpu(cpu, &conn_allowed_cpumask) {
 		if (ord-- == 0) {
 			cpumask_set_cpu(cpu, conn->conn_cpumask);
 			return;
@@ -3609,6 +3624,62 @@ void iscsit_thread_get_cpumask(struct iscsi_conn *conn)
 	cpumask_setall(conn->conn_cpumask);
 }
 
+static void iscsit_thread_reschedule(struct iscsi_conn *conn)
+{
+	/*
+	 * If iscsit_global->allowed_cpumask modified, reschedule iSCSI
+	 * connection's RX/TX threads update conn->allowed_cpumask.
+	 */
+	if (!cpumask_equal(iscsit_global->allowed_cpumask,
+			   conn->allowed_cpumask)) {
+		iscsit_thread_get_cpumask(conn);
+		conn->conn_tx_reset_cpumask = 1;
+		conn->conn_rx_reset_cpumask = 1;
+		cpumask_copy(conn->allowed_cpumask,
+			     iscsit_global->allowed_cpumask);
+	}
+}
+
+void iscsit_thread_check_cpumask(
+	struct iscsi_conn *conn,
+	struct task_struct *p,
+	int mode)
+{
+	/*
+	 * The TX and RX threads maybe call iscsit_thread_check_cpumask()
+	 * at the same time. The RX thread might be faster and return from
+	 * iscsit_thread_reschedule() with conn_rx_reset_cpumask set to 0.
+	 * Then the TX thread sets it back to 1.
+	 * The next time the RX thread loops, it sees conn_rx_reset_cpumask
+	 * set to 1 and calls set_cpus_allowed_ptr() again and set it to 0.
+	 */
+	iscsit_thread_reschedule(conn);
+
+	/*
+	 * mode == 1 signals iscsi_target_tx_thread() usage.
+	 * mode == 0 signals iscsi_target_rx_thread() usage.
+	 */
+	if (mode == 1) {
+		if (!conn->conn_tx_reset_cpumask)
+			return;
+	} else {
+		if (!conn->conn_rx_reset_cpumask)
+			return;
+	}
+
+	/*
+	 * Update the CPU mask for this single kthread so that
+	 * both TX and RX kthreads are scheduled to run on the
+	 * same CPU.
+	 */
+	set_cpus_allowed_ptr(p, conn->conn_cpumask);
+	if (mode == 1)
+		conn->conn_tx_reset_cpumask = 0;
+	else
+		conn->conn_rx_reset_cpumask = 0;
+}
+EXPORT_SYMBOL(iscsit_thread_check_cpumask);
+
 int
 iscsit_immediate_queue(struct iscsi_conn *conn, struct iscsi_cmd *cmd, int state)
 {
diff --git a/drivers/target/iscsi/iscsi_target_configfs.c b/drivers/target/iscsi/iscsi_target_configfs.c
index 2a9de24a8bbe2..0cedcfe207b56 100644
--- a/drivers/target/iscsi/iscsi_target_configfs.c
+++ b/drivers/target/iscsi/iscsi_target_configfs.c
@@ -1127,8 +1127,40 @@ static ssize_t lio_target_wwn_lio_version_show(struct config_item *item,
 
 CONFIGFS_ATTR_RO(lio_target_wwn_, lio_version);
 
+static ssize_t lio_target_wwn_cpus_allowed_list_show(
+		struct config_item *item, char *page)
+{
+	return sprintf(page, "%*pbl\n",
+		       cpumask_pr_args(iscsit_global->allowed_cpumask));
+}
+
+static ssize_t lio_target_wwn_cpus_allowed_list_store(
+		struct config_item *item, const char *page, size_t count)
+{
+	int ret;
+	char *orig;
+	cpumask_t new_allowed_cpumask;
+
+	orig = kstrdup(page, GFP_KERNEL);
+	if (!orig)
+		return -ENOMEM;
+
+	cpumask_clear(&new_allowed_cpumask);
+	ret = cpulist_parse(orig, &new_allowed_cpumask);
+
+	kfree(orig);
+	if (ret != 0)
+		return ret;
+
+	cpumask_copy(iscsit_global->allowed_cpumask, &new_allowed_cpumask);
+	return count;
+}
+
+CONFIGFS_ATTR(lio_target_wwn_, cpus_allowed_list);
+
 static struct configfs_attribute *lio_target_wwn_attrs[] = {
 	&lio_target_wwn_attr_lio_version,
+	&lio_target_wwn_attr_cpus_allowed_list,
 	NULL,
 };
 
diff --git a/drivers/target/iscsi/iscsi_target_login.c b/drivers/target/iscsi/iscsi_target_login.c
index 1a9c50401bdb5..9c01fb8645853 100644
--- a/drivers/target/iscsi/iscsi_target_login.c
+++ b/drivers/target/iscsi/iscsi_target_login.c
@@ -1129,8 +1129,15 @@ static struct iscsi_conn *iscsit_alloc_conn(struct iscsi_np *np)
 		goto free_conn_ops;
 	}
 
+	if (!zalloc_cpumask_var(&conn->allowed_cpumask, GFP_KERNEL)) {
+		pr_err("Unable to allocate conn->allowed_cpumask\n");
+		goto free_conn_cpumask;
+	}
+
 	return conn;
 
+free_conn_cpumask:
+	free_cpumask_var(conn->conn_cpumask);
 free_conn_ops:
 	kfree(conn->conn_ops);
 put_transport:
@@ -1142,6 +1149,7 @@ free_conn:
 
 void iscsit_free_conn(struct iscsi_conn *conn)
 {
+	free_cpumask_var(conn->allowed_cpumask);
 	free_cpumask_var(conn->conn_cpumask);
 	kfree(conn->conn_ops);
 	iscsit_put_transport(conn->conn_transport);