1 files changed, 201 insertions, 152 deletions

diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index f0540c557515..42ee7fce3d95 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -56,7 +56,6 @@ static bool init_net_initialized;
  * outside.
  */
 DECLARE_RWSEM(pernet_ops_rwsem);
-EXPORT_SYMBOL_GPL(pernet_ops_rwsem);
 
 #define MIN_PERNET_OPS_ID	\
 	((sizeof(struct net_generic) + sizeof(void *) - 1) / sizeof(void *))
@@ -69,12 +68,15 @@ DEFINE_COOKIE(net_cookie);
 
 static struct net_generic *net_alloc_generic(void)
 {
+	unsigned int gen_ptrs = READ_ONCE(max_gen_ptrs);
+	unsigned int generic_size;
 	struct net_generic *ng;
-	unsigned int generic_size = offsetof(struct net_generic, ptr[max_gen_ptrs]);
+
+	generic_size = offsetof(struct net_generic, ptr[gen_ptrs]);
 
 	ng = kzalloc(generic_size, GFP_KERNEL);
 	if (ng)
-		ng->s.len = max_gen_ptrs;
+		ng->s.len = gen_ptrs;
 
 	return ng;
 }
@@ -122,7 +124,7 @@ static int ops_init(const struct pernet_operations *ops, struct net *net)
 	int err = -ENOMEM;
 	void *data = NULL;
 
-	if (ops->id && ops->size) {
+	if (ops->id) {
 		data = kzalloc(ops->size, GFP_KERNEL);
 		if (!data)
 			goto out;
@@ -137,7 +139,7 @@ static int ops_init(const struct pernet_operations *ops, struct net *net)
 	if (!err)
 		return 0;
 
-	if (ops->id && ops->size) {
+	if (ops->id) {
 		ng = rcu_dereference_protected(net->gen,
 					       lockdep_is_held(&pernet_ops_rwsem));
 		ng->ptr[*ops->id] = NULL;
@@ -161,16 +163,45 @@ static void ops_pre_exit_list(const struct pernet_operations *ops,
 	}
 }
 
+static void ops_exit_rtnl_list(const struct list_head *ops_list,
+			       const struct pernet_operations *ops,
+			       struct list_head *net_exit_list)
+{
+	const struct pernet_operations *saved_ops = ops;
+	LIST_HEAD(dev_kill_list);
+	struct net *net;
+
+	rtnl_lock();
+
+	list_for_each_entry(net, net_exit_list, exit_list) {
+		__rtnl_net_lock(net);
+
+		ops = saved_ops;
+		list_for_each_entry_continue_reverse(ops, ops_list, list) {
+			if (ops->exit_rtnl)
+				ops->exit_rtnl(net, &dev_kill_list);
+		}
+
+		__rtnl_net_unlock(net);
+	}
+
+	unregister_netdevice_many(&dev_kill_list);
+
+	rtnl_unlock();
+}
+
 static void ops_exit_list(const struct pernet_operations *ops,
 			  struct list_head *net_exit_list)
 {
-	struct net *net;
 	if (ops->exit) {
+		struct net *net;
+
 		list_for_each_entry(net, net_exit_list, exit_list) {
 			ops->exit(net);
 			cond_resched();
 		}
 	}
+
 	if (ops->exit_batch)
 		ops->exit_batch(net_exit_list);
 }
@@ -179,12 +210,63 @@ static void ops_free_list(const struct pernet_operations *ops,
 			  struct list_head *net_exit_list)
 {
 	struct net *net;
-	if (ops->size && ops->id) {
+
+	if (ops->id) {
 		list_for_each_entry(net, net_exit_list, exit_list)
 			kfree(net_generic(net, *ops->id));
 	}
 }
 
+static void ops_undo_list(const struct list_head *ops_list,
+			  const struct pernet_operations *ops,
+			  struct list_head *net_exit_list,
+			  bool expedite_rcu)
+{
+	const struct pernet_operations *saved_ops;
+	bool hold_rtnl = false;
+
+	if (!ops)
+		ops = list_entry(ops_list, typeof(*ops), list);
+
+	saved_ops = ops;
+
+	list_for_each_entry_continue_reverse(ops, ops_list, list) {
+		hold_rtnl |= !!ops->exit_rtnl;
+		ops_pre_exit_list(ops, net_exit_list);
+	}
+
+	/* Another CPU might be rcu-iterating the list, wait for it.
+	 * This needs to be before calling the exit() notifiers, so the
+	 * rcu_barrier() after ops_undo_list() isn't sufficient alone.
+	 * Also the pre_exit() and exit() methods need this barrier.
+	 */
+	if (expedite_rcu)
+		synchronize_rcu_expedited();
+	else
+		synchronize_rcu();
+
+	if (hold_rtnl)
+		ops_exit_rtnl_list(ops_list, saved_ops, net_exit_list);
+
+	ops = saved_ops;
+	list_for_each_entry_continue_reverse(ops, ops_list, list)
+		ops_exit_list(ops, net_exit_list);
+
+	ops = saved_ops;
+	list_for_each_entry_continue_reverse(ops, ops_list, list)
+		ops_free_list(ops, net_exit_list);
+}
+
+static void ops_undo_single(struct pernet_operations *ops,
+			    struct list_head *net_exit_list)
+{
+	LIST_HEAD(ops_list);
+
+	list_add(&ops->list, &ops_list);
+	ops_undo_list(&ops_list, NULL, net_exit_list, false);
+	list_del(&ops->list);
+}
+
 /* should be called with nsid_lock held */
 static int alloc_netid(struct net *net, struct net *peer, int reqid)
 {
@@ -305,36 +387,56 @@ struct net *get_net_ns_by_id(const struct net *net, int id)
 }
 EXPORT_SYMBOL_GPL(get_net_ns_by_id);
 
+static __net_init void preinit_net_sysctl(struct net *net)
+{
+	net->core.sysctl_somaxconn = SOMAXCONN;
+	/* Limits per socket sk_omem_alloc usage.
+	 * TCP zerocopy regular usage needs 128 KB.
+	 */
+	net->core.sysctl_optmem_max = 128 * 1024;
+	net->core.sysctl_txrehash = SOCK_TXREHASH_ENABLED;
+	net->core.sysctl_tstamp_allow_data = 1;
+}
+
 /* init code that must occur even if setup_net() is not called. */
-static __net_init void preinit_net(struct net *net)
+static __net_init void preinit_net(struct net *net, struct user_namespace *user_ns)
 {
+	refcount_set(&net->passive, 1);
+	refcount_set(&net->ns.count, 1);
+	ref_tracker_dir_init(&net->refcnt_tracker, 128, "net refcnt");
 	ref_tracker_dir_init(&net->notrefcnt_tracker, 128, "net notrefcnt");
+
+	get_random_bytes(&net->hash_mix, sizeof(u32));
+	net->dev_base_seq = 1;
+	net->user_ns = user_ns;
+
+	idr_init(&net->netns_ids);
+	spin_lock_init(&net->nsid_lock);
+	mutex_init(&net->ipv4.ra_mutex);
+
+#ifdef CONFIG_DEBUG_NET_SMALL_RTNL
+	mutex_init(&net->rtnl_mutex);
+	lock_set_cmp_fn(&net->rtnl_mutex, rtnl_net_lock_cmp_fn, NULL);
+#endif
+
+	INIT_LIST_HEAD(&net->ptype_all);
+	INIT_LIST_HEAD(&net->ptype_specific);
+	preinit_net_sysctl(net);
 }
 
 /*
  * setup_net runs the initializers for the network namespace object.
  */
-static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
+static __net_init int setup_net(struct net *net)
 {
 	/* Must be called with pernet_ops_rwsem held */
-	const struct pernet_operations *ops, *saved_ops;
+	const struct pernet_operations *ops;
 	LIST_HEAD(net_exit_list);
-	LIST_HEAD(dev_kill_list);
 	int error = 0;
 
-	refcount_set(&net->ns.count, 1);
-	ref_tracker_dir_init(&net->refcnt_tracker, 128, "net refcnt");
-
-	refcount_set(&net->passive, 1);
-	get_random_bytes(&net->hash_mix, sizeof(u32));
 	preempt_disable();
 	net->net_cookie = gen_cookie_next(&net_cookie);
 	preempt_enable();
-	net->dev_base_seq = 1;
-	net->user_ns = user_ns;
-	idr_init(&net->netns_ids);
-	spin_lock_init(&net->nsid_lock);
-	mutex_init(&net->ipv4.ra_mutex);
 
 	list_for_each_entry(ops, &pernet_list, list) {
 		error = ops_init(ops, net);
@@ -352,59 +454,11 @@ out_undo:
 	 * for the pernet modules whose init functions did not fail.
 	 */
 	list_add(&net->exit_list, &net_exit_list);
-	saved_ops = ops;
-	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
-		ops_pre_exit_list(ops, &net_exit_list);
-
-	synchronize_rcu();
-
-	ops = saved_ops;
-	rtnl_lock();
-	list_for_each_entry_continue_reverse(ops, &pernet_list, list) {
-		if (ops->exit_batch_rtnl)
-			ops->exit_batch_rtnl(&net_exit_list, &dev_kill_list);
-	}
-	unregister_netdevice_many(&dev_kill_list);
-	rtnl_unlock();
-
-	ops = saved_ops;
-	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
-		ops_exit_list(ops, &net_exit_list);
-
-	ops = saved_ops;
-	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
-		ops_free_list(ops, &net_exit_list);
-
+	ops_undo_list(&pernet_list, ops, &net_exit_list, false);
 	rcu_barrier();
 	goto out;
 }
 
-static int __net_init net_defaults_init_net(struct net *net)
-{
-	net->core.sysctl_somaxconn = SOMAXCONN;
-	/* Limits per socket sk_omem_alloc usage.
-	 * TCP zerocopy regular usage needs 128 KB.
-	 */
-	net->core.sysctl_optmem_max = 128 * 1024;
-	net->core.sysctl_txrehash = SOCK_TXREHASH_ENABLED;
-
-	return 0;
-}
-
-static struct pernet_operations net_defaults_ops = {
-	.init = net_defaults_init_net,
-};
-
-static __init int net_defaults_init(void)
-{
-	if (register_pernet_subsys(&net_defaults_ops))
-		panic("Cannot initialize net default settings");
-
-	return 0;
-}
-
-core_initcall(net_defaults_init);
-
 #ifdef CONFIG_NET_NS
 static struct ucounts *inc_net_namespaces(struct user_namespace *ns)
 {
@@ -453,7 +507,22 @@ out_free:
 	goto out;
 }
 
-static void net_free(struct net *net)
+static LLIST_HEAD(defer_free_list);
+
+static void net_complete_free(void)
+{
+	struct llist_node *kill_list;
+	struct net *net, *next;
+
+	/* Get the list of namespaces to free from last round. */
+	kill_list = llist_del_all(&defer_free_list);
+
+	llist_for_each_entry_safe(net, next, kill_list, defer_free_list)
+		kmem_cache_free(net_cachep, net);
+
+}
+
+void net_passive_dec(struct net *net)
 {
 	if (refcount_dec_and_test(&net->passive)) {
 		kfree(rcu_access_pointer(net->gen));
@@ -461,7 +530,8 @@ static void net_free(struct net *net)
 		/* There should not be any trackers left there. */
 		ref_tracker_dir_exit(&net->notrefcnt_tracker);
 
-		kmem_cache_free(net_cachep, net);
+		/* Wait for an extra rcu_barrier() before final free. */
+		llist_add(&net->defer_free_list, &defer_free_list);
 	}
 }
 
@@ -470,7 +540,7 @@ void net_drop_ns(void *p)
 	struct net *net = (struct net *)p;
 
 	if (net)
-		net_free(net);
+		net_passive_dec(net);
 }
 
 struct net *copy_net_ns(unsigned long flags,
@@ -493,8 +563,7 @@ struct net *copy_net_ns(unsigned long flags,
 		goto dec_ucounts;
 	}
 
-	preinit_net(net);
-	refcount_set(&net->passive, 1);
+	preinit_net(net, user_ns);
 	net->ucounts = ucounts;
 	get_user_ns(user_ns);
 
@@ -502,7 +571,7 @@ struct net *copy_net_ns(unsigned long flags,
 	if (rv < 0)
 		goto put_userns;
 
-	rv = setup_net(net, user_ns);
+	rv = setup_net(net);
 
 	up_read(&pernet_ops_rwsem);
 
@@ -512,7 +581,7 @@ put_userns:
 		key_remove_domain(net->key_domain);
 #endif
 		put_user_ns(user_ns);
-		net_free(net);
+		net_passive_dec(net);
 dec_ucounts:
 		dec_net_namespaces(ucounts);
 		return ERR_PTR(rv);
@@ -577,13 +646,15 @@ static void unhash_nsid(struct net *net, struct net *last)
 
 static LLIST_HEAD(cleanup_list);
 
+struct task_struct *cleanup_net_task;
+
 static void cleanup_net(struct work_struct *work)
 {
-	const struct pernet_operations *ops;
-	struct net *net, *tmp, *last;
 	struct llist_node *net_kill_list;
+	struct net *net, *tmp, *last;
 	LIST_HEAD(net_exit_list);
-	LIST_HEAD(dev_kill_list);
+
+	cleanup_net_task = current;
 
 	/* Atomically snapshot the list of namespaces to cleanup */
 	net_kill_list = llist_del_all(&cleanup_list);
@@ -612,33 +683,7 @@ static void cleanup_net(struct work_struct *work)
 		list_add_tail(&net->exit_list, &net_exit_list);
 	}
 
-	/* Run all of the network namespace pre_exit methods */
-	list_for_each_entry_reverse(ops, &pernet_list, list)
-		ops_pre_exit_list(ops, &net_exit_list);
-
-	/*
-	 * Another CPU might be rcu-iterating the list, wait for it.
-	 * This needs to be before calling the exit() notifiers, so
-	 * the rcu_barrier() below isn't sufficient alone.
-	 * Also the pre_exit() and exit() methods need this barrier.
-	 */
-	synchronize_rcu_expedited();
-
-	rtnl_lock();
-	list_for_each_entry_reverse(ops, &pernet_list, list) {
-		if (ops->exit_batch_rtnl)
-			ops->exit_batch_rtnl(&net_exit_list, &dev_kill_list);
-	}
-	unregister_netdevice_many(&dev_kill_list);
-	rtnl_unlock();
-
-	/* Run all of the network namespace exit methods */
-	list_for_each_entry_reverse(ops, &pernet_list, list)
-		ops_exit_list(ops, &net_exit_list);
-
-	/* Free the net generic variables */
-	list_for_each_entry_reverse(ops, &pernet_list, list)
-		ops_free_list(ops, &net_exit_list);
+	ops_undo_list(&pernet_list, NULL, &net_exit_list, true);
 
 	up_read(&pernet_ops_rwsem);
 
@@ -647,6 +692,8 @@ static void cleanup_net(struct work_struct *work)
 	 */
 	rcu_barrier();
 
+	net_complete_free();
+
 	/* Finally it is safe to free my network namespace structure */
 	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
 		list_del_init(&net->exit_list);
@@ -655,8 +702,9 @@ static void cleanup_net(struct work_struct *work)
 		key_remove_domain(net->key_domain);
 #endif
 		put_user_ns(net->user_ns);
-		net_free(net);
+		net_passive_dec(net);
 	}
+	cleanup_net_task = NULL;
 }
 
 /**
@@ -690,30 +738,33 @@ EXPORT_SYMBOL_GPL(__put_net);
  * get_net_ns - increment the refcount of the network namespace
  * @ns: common namespace (net)
  *
- * Returns the net's common namespace.
+ * Returns the net's common namespace or ERR_PTR() if ref is zero.
  */
 struct ns_common *get_net_ns(struct ns_common *ns)
 {
-	return &get_net(container_of(ns, struct net, ns))->ns;
+	struct net *net;
+
+	net = maybe_get_net(container_of(ns, struct net, ns));
+	if (net)
+		return &net->ns;
+	return ERR_PTR(-EINVAL);
 }
 EXPORT_SYMBOL_GPL(get_net_ns);
 
 struct net *get_net_ns_by_fd(int fd)
 {
-	struct fd f = fdget(fd);
-	struct net *net = ERR_PTR(-EINVAL);
+	CLASS(fd, f)(fd);
 
-	if (!f.file)
+	if (fd_empty(f))
 		return ERR_PTR(-EBADF);
 
-	if (proc_ns_file(f.file)) {
-		struct ns_common *ns = get_proc_ns(file_inode(f.file));
+	if (proc_ns_file(fd_file(f))) {
+		struct ns_common *ns = get_proc_ns(file_inode(fd_file(f)));
 		if (ns->ops == &netns_operations)
-			net = get_net(container_of(ns, struct net, ns));
+			return get_net(container_of(ns, struct net, ns));
 	}
-	fdput(f);
 
-	return net;
+	return ERR_PTR(-EINVAL);
 }
 EXPORT_SYMBOL_GPL(get_net_ns_by_fd);
 #endif
@@ -1090,7 +1141,7 @@ static int rtnl_net_dumpid(struct sk_buff *skb, struct netlink_callback *cb)
 end:
 	if (net_cb.fillargs.add_ref)
 		put_net(net_cb.tgt_net);
-	return err < 0 ? err : skb->len;
+	return err;
 }
 
 static void rtnl_net_notifyid(struct net *net, int cmd, int id, u32 portid,
@@ -1159,13 +1210,23 @@ static void __init netns_ipv4_struct_check(void)
 	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
 				      sysctl_tcp_early_demux);
 	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
+				      sysctl_tcp_l3mdev_accept);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
 				      sysctl_tcp_reordering);
 	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
 				      sysctl_tcp_rmem);
-	CACHELINE_ASSERT_GROUP_SIZE(struct netns_ipv4, netns_ipv4_read_rx, 18);
+	CACHELINE_ASSERT_GROUP_SIZE(struct netns_ipv4, netns_ipv4_read_rx, 22);
 }
 #endif
 
+static const struct rtnl_msg_handler net_ns_rtnl_msg_handlers[] __initconst = {
+	{.msgtype = RTM_NEWNSID, .doit = rtnl_net_newid,
+	 .flags = RTNL_FLAG_DOIT_UNLOCKED},
+	{.msgtype = RTM_GETNSID, .doit = rtnl_net_getid,
+	 .dumpit = rtnl_net_dumpid,
+	 .flags = RTNL_FLAG_DOIT_UNLOCKED | RTNL_FLAG_DUMP_UNLOCKED},
+};
+
 void __init net_ns_init(void)
 {
 	struct net_generic *ng;
@@ -1191,9 +1252,10 @@ void __init net_ns_init(void)
 #ifdef CONFIG_KEYS
 	init_net.key_domain = &init_net_key_domain;
 #endif
+	preinit_net(&init_net, &init_user_ns);
+
 	down_write(&pernet_ops_rwsem);
-	preinit_net(&init_net);
-	if (setup_net(&init_net, &init_user_ns))
+	if (setup_net(&init_net))
 		panic("Could not setup the initial network namespace");
 
 	init_net_initialized = true;
@@ -1202,40 +1264,19 @@ void __init net_ns_init(void)
 	if (register_pernet_subsys(&net_ns_ops))
 		panic("Could not register network namespace subsystems");
 
-	rtnl_register(PF_UNSPEC, RTM_NEWNSID, rtnl_net_newid, NULL,
-		      RTNL_FLAG_DOIT_UNLOCKED);
-	rtnl_register(PF_UNSPEC, RTM_GETNSID, rtnl_net_getid, rtnl_net_dumpid,
-		      RTNL_FLAG_DOIT_UNLOCKED);
-}
-
-static void free_exit_list(struct pernet_operations *ops, struct list_head *net_exit_list)
-{
-	ops_pre_exit_list(ops, net_exit_list);
-	synchronize_rcu();
-
-	if (ops->exit_batch_rtnl) {
-		LIST_HEAD(dev_kill_list);
-
-		rtnl_lock();
-		ops->exit_batch_rtnl(net_exit_list, &dev_kill_list);
-		unregister_netdevice_many(&dev_kill_list);
-		rtnl_unlock();
-	}
-	ops_exit_list(ops, net_exit_list);
-
-	ops_free_list(ops, net_exit_list);
+	rtnl_register_many(net_ns_rtnl_msg_handlers);
 }
 
 #ifdef CONFIG_NET_NS
 static int __register_pernet_operations(struct list_head *list,
 					struct pernet_operations *ops)
 {
+	LIST_HEAD(net_exit_list);
 	struct net *net;
 	int error;
-	LIST_HEAD(net_exit_list);
 
 	list_add_tail(&ops->list, list);
-	if (ops->init || (ops->id && ops->size)) {
+	if (ops->init || ops->id) {
 		/* We held write locked pernet_ops_rwsem, and parallel
 		 * setup_net() and cleanup_net() are not possible.
 		 */
@@ -1251,21 +1292,21 @@ static int __register_pernet_operations(struct list_head *list,
 out_undo:
 	/* If I have an error cleanup all namespaces I initialized */
 	list_del(&ops->list);
-	free_exit_list(ops, &net_exit_list);
+	ops_undo_single(ops, &net_exit_list);
 	return error;
 }
 
 static void __unregister_pernet_operations(struct pernet_operations *ops)
 {
-	struct net *net;
 	LIST_HEAD(net_exit_list);
+	struct net *net;
 
-	list_del(&ops->list);
 	/* See comment in __register_pernet_operations() */
 	for_each_net(net)
 		list_add_tail(&net->exit_list, &net_exit_list);
 
-	free_exit_list(ops, &net_exit_list);
+	list_del(&ops->list);
+	ops_undo_single(ops, &net_exit_list);
 }
 
 #else
@@ -1287,8 +1328,9 @@ static void __unregister_pernet_operations(struct pernet_operations *ops)
 		list_del(&ops->list);
 	} else {
 		LIST_HEAD(net_exit_list);
+
 		list_add(&init_net.exit_list, &net_exit_list);
-		free_exit_list(ops, &net_exit_list);
+		ops_undo_single(ops, &net_exit_list);
 	}
 }
 
@@ -1301,13 +1343,20 @@ static int register_pernet_operations(struct list_head *list,
 {
 	int error;
 
+	if (WARN_ON(!!ops->id ^ !!ops->size))
+		return -EINVAL;
+
 	if (ops->id) {
 		error = ida_alloc_min(&net_generic_ids, MIN_PERNET_OPS_ID,
 				GFP_KERNEL);
 		if (error < 0)
 			return error;
 		*ops->id = error;
-		max_gen_ptrs = max(max_gen_ptrs, *ops->id + 1);
+		/* This does not require READ_ONCE as writers already hold
+		 * pernet_ops_rwsem. But WRITE_ONCE is needed to protect
+		 * net_alloc_generic.
+		 */
+		WRITE_ONCE(max_gen_ptrs, max(max_gen_ptrs, *ops->id + 1));
 	}
 	error = __register_pernet_operations(list, ops);
 	if (error) {