Merge branch 'net-Get-rid-of-net_mutex-and-simplify-cleanup_list-queueing'

Kirill Tkhai says:

====================
net: Get rid of net_mutex and simplify cleanup_list queueing

[1/3] kills net_mutex and makes net_sem be taken for write instead.
      This is made to take less locks (1 instead of 2) for the time
      before all pernet_operations are converted.

[2-3/3] simplifies dead net cleanup queueing, and makes llist api
        be used for that.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

include/linux/rtnetlink.h

@@ -35,7 +35,6 @@ extern int rtnl_trylock(void);
 extern int rtnl_is_locked(void);
 
 extern wait_queue_head_t netdev_unregistering_wq;
-extern struct mutex net_mutex;
 extern struct rw_semaphore net_sem;
 
 #ifdef CONFIG_PROVE_LOCKING

include/net/net_namespace.h

@@ -59,8 +59,12 @@ struct net {
 	atomic64_t		cookie_gen;
 
 	struct list_head	list;		/* list of network namespaces */
-	struct list_head	cleanup_list;	/* namespaces on death row */
-	struct list_head	exit_list;	/* Use only net_mutex */
+	struct list_head	exit_list;	/* To linked to call pernet exit
+						 * methods on dead net (net_sem
+						 * read locked), or to unregister
+						 * pernet ops (net_sem wr locked).
+						 */
+	struct llist_node	cleanup_list;	/* namespaces on death row */
 
 	struct user_namespace   *user_ns;	/* Owning user namespace */
 	struct ucounts		*ucounts;
@@ -89,7 +93,7 @@ struct net {
 	/* core fib_rules */
 	struct list_head	rules_ops;
 
-	struct list_head	fib_notifier_ops;  /* protected by net_mutex */
+	struct list_head	fib_notifier_ops;  /* protected by net_sem */
 
 	struct net_device       *loopback_dev;          /* The loopback */
 	struct netns_core	core;
@@ -316,7 +320,7 @@ struct pernet_operations {
 	/*
 	 * Indicates above methods are allowed to be executed in parallel
 	 * with methods of any other pernet_operations, i.e. they are not
-	 * need synchronization via net_mutex.
+	 * need write locked net_sem.
 	 */
 	bool async;
 };

net/core/net_namespace.c

@@ -29,8 +29,6 @@
 
 static LIST_HEAD(pernet_list);
 static struct list_head *first_device = &pernet_list;
-/* Used only if there are !async pernet_operations registered */
-DEFINE_MUTEX(net_mutex);
 
 LIST_HEAD(net_namespace_list);
 EXPORT_SYMBOL_GPL(net_namespace_list);
@@ -407,6 +405,7 @@ struct net *copy_net_ns(unsigned long flags,
 {
 	struct ucounts *ucounts;
 	struct net *net;
+	unsigned write;
 	int rv;
 
 	if (!(flags & CLONE_NEWNET))
@@ -424,20 +423,26 @@ struct net *copy_net_ns(unsigned long flags,
 	refcount_set(&net->passive, 1);
 	net->ucounts = ucounts;
 	get_user_ns(user_ns);
-
-	rv = down_read_killable(&net_sem);
+again:
+	write = READ_ONCE(nr_sync_pernet_ops);
+	if (write)
+		rv = down_write_killable(&net_sem);
+	else
+		rv = down_read_killable(&net_sem);
 	if (rv < 0)
 		goto put_userns;
-	if (nr_sync_pernet_ops) {
-		rv = mutex_lock_killable(&net_mutex);
-		if (rv < 0)
-			goto up_read;
+
+	if (!write && unlikely(READ_ONCE(nr_sync_pernet_ops))) {
+		up_read(&net_sem);
+		goto again;
 	}
 	rv = setup_net(net, user_ns);
-	if (nr_sync_pernet_ops)
-		mutex_unlock(&net_mutex);
-up_read:
-	up_read(&net_sem);
+
+	if (write)
+		up_write(&net_sem);
+	else
+		up_read(&net_sem);
+
 	if (rv < 0) {
 put_userns:
 		put_user_ns(user_ns);
@@ -476,28 +481,33 @@ static void unhash_nsid(struct net *net, struct net *last)
 	spin_unlock_bh(&net->nsid_lock);
 }
 
-static DEFINE_SPINLOCK(cleanup_list_lock);
-static LIST_HEAD(cleanup_list);  /* Must hold cleanup_list_lock to touch */
+static LLIST_HEAD(cleanup_list);
 
 static void cleanup_net(struct work_struct *work)
 {
 	const struct pernet_operations *ops;
 	struct net *net, *tmp, *last;
-	struct list_head net_kill_list;
+	struct llist_node *net_kill_list;
 	LIST_HEAD(net_exit_list);
+	unsigned write;
 
 	/* Atomically snapshot the list of namespaces to cleanup */
-	spin_lock_irq(&cleanup_list_lock);
-	list_replace_init(&cleanup_list, &net_kill_list);
-	spin_unlock_irq(&cleanup_list_lock);
+	net_kill_list = llist_del_all(&cleanup_list);
+again:
+	write = READ_ONCE(nr_sync_pernet_ops);
+	if (write)
+		down_write(&net_sem);
+	else
+		down_read(&net_sem);
 
-	down_read(&net_sem);
-	if (nr_sync_pernet_ops)
-		mutex_lock(&net_mutex);
+	if (!write && unlikely(READ_ONCE(nr_sync_pernet_ops))) {
+		up_read(&net_sem);
+		goto again;
+	}
 
 	/* Don't let anyone else find us. */
 	rtnl_lock();
-	list_for_each_entry(net, &net_kill_list, cleanup_list)
+	llist_for_each_entry(net, net_kill_list, cleanup_list)
 		list_del_rcu(&net->list);
 	/* Cache last net. After we unlock rtnl, no one new net
 	 * added to net_namespace_list can assign nsid pointer
@@ -512,7 +522,7 @@ static void cleanup_net(struct work_struct *work)
 	last = list_last_entry(&net_namespace_list, struct net, list);
 	rtnl_unlock();
 
-	list_for_each_entry(net, &net_kill_list, cleanup_list) {
+	llist_for_each_entry(net, net_kill_list, cleanup_list) {
 		unhash_nsid(net, last);
 		list_add_tail(&net->exit_list, &net_exit_list);
 	}
@@ -528,14 +538,14 @@ static void cleanup_net(struct work_struct *work)
 	list_for_each_entry_reverse(ops, &pernet_list, list)
 		ops_exit_list(ops, &net_exit_list);
 
-	if (nr_sync_pernet_ops)
-		mutex_unlock(&net_mutex);
-
 	/* Free the net generic variables */
 	list_for_each_entry_reverse(ops, &pernet_list, list)
 		ops_free_list(ops, &net_exit_list);
 
-	up_read(&net_sem);
+	if (write)
+		up_write(&net_sem);
+	else
+		up_read(&net_sem);
 
 	/* Ensure there are no outstanding rcu callbacks using this
 	 * network namespace.
@@ -563,8 +573,6 @@ static void cleanup_net(struct work_struct *work)
 void net_ns_barrier(void)
 {
 	down_write(&net_sem);
-	mutex_lock(&net_mutex);
-	mutex_unlock(&net_mutex);
 	up_write(&net_sem);
 }
 EXPORT_SYMBOL(net_ns_barrier);
@@ -574,13 +582,8 @@ static DECLARE_WORK(net_cleanup_work, cleanup_net);
 void __put_net(struct net *net)
 {
 	/* Cleanup the network namespace in process context */
-	unsigned long flags;
-
-	spin_lock_irqsave(&cleanup_list_lock, flags);
-	list_add(&net->cleanup_list, &cleanup_list);
-	spin_unlock_irqrestore(&cleanup_list_lock, flags);
-
-	queue_work(netns_wq, &net_cleanup_work);
+	if (llist_add(&net->cleanup_list, &cleanup_list))
+		queue_work(netns_wq, &net_cleanup_work);
 }
 EXPORT_SYMBOL_GPL(__put_net);