1 files changed, 436 insertions, 211 deletions

diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index 17f36317363d..a6e6a964a287 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -1,3 +1,4 @@
+// SPDX-License-Identifier: GPL-2.0-only
 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 
 #include <linux/workqueue.h>
@@ -18,7 +19,10 @@
 #include <linux/net_namespace.h>
 #include <linux/sched/task.h>
 #include <linux/uidgid.h>
+#include <linux/proc_fs.h>
+#include <linux/nstree.h>
 
+#include <net/aligned_data.h>
 #include <net/sock.h>
 #include <net/netlink.h>
 #include <net/net_namespace.h>
@@ -38,10 +42,11 @@ EXPORT_SYMBOL_GPL(net_namespace_list);
 DECLARE_RWSEM(net_rwsem);
 EXPORT_SYMBOL_GPL(net_rwsem);
 
-struct net init_net = {
-	.count		= REFCOUNT_INIT(1),
-	.dev_base_head	= LIST_HEAD_INIT(init_net.dev_base_head),
-};
+#ifdef CONFIG_KEYS
+static struct key_tag init_net_key_domain = { .usage = REFCOUNT_INIT(1) };
+#endif
+
+struct net init_net;
 EXPORT_SYMBOL(init_net);
 
 static bool init_net_initialized;
@@ -52,7 +57,6 @@ 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 *))
@@ -63,12 +67,15 @@ static unsigned int max_gen_ptrs = INITIAL_NET_GEN_PTRS;
 
 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;
 }
@@ -87,7 +94,7 @@ static int net_assign_generic(struct net *net, unsigned int id, void *data)
 	}
 
 	ng = net_alloc_generic();
-	if (ng == NULL)
+	if (!ng)
 		return -ENOMEM;
 
 	/*
@@ -112,10 +119,11 @@ static int net_assign_generic(struct net *net, unsigned int id, void *data)
 
 static int ops_init(const struct pernet_operations *ops, struct net *net)
 {
+	struct net_generic *ng;
 	int err = -ENOMEM;
 	void *data = NULL;
 
-	if (ops->id && ops->size) {
+	if (ops->id) {
 		data = kzalloc(ops->size, GFP_KERNEL);
 		if (!data)
 			goto out;
@@ -130,6 +138,12 @@ static int ops_init(const struct pernet_operations *ops, struct net *net)
 	if (!err)
 		return 0;
 
+	if (ops->id) {
+		ng = rcu_dereference_protected(net->gen,
+					       lockdep_is_held(&pernet_ops_rwsem));
+		ng->ptr[*ops->id] = NULL;
+	}
+
 cleanup:
 	kfree(data);
 
@@ -137,21 +151,56 @@ out:
 	return err;
 }
 
-static void ops_free(const struct pernet_operations *ops, struct net *net)
+static void ops_pre_exit_list(const struct pernet_operations *ops,
+			      struct list_head *net_exit_list)
+{
+	struct net *net;
+
+	if (ops->pre_exit) {
+		list_for_each_entry(net, net_exit_list, exit_list)
+			ops->pre_exit(net);
+	}
+}
+
+static void ops_exit_rtnl_list(const struct list_head *ops_list,
+			       const struct pernet_operations *ops,
+			       struct list_head *net_exit_list)
 {
-	if (ops->id && ops->size) {
-		kfree(net_generic(net, *ops->id));
+	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) {
-		list_for_each_entry(net, net_exit_list, exit_list)
+		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);
 }
@@ -160,10 +209,61 @@ 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)
-			ops_free(ops, net);
+			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 */
@@ -192,16 +292,10 @@ static int net_eq_idr(int id, void *net, void *peer)
 	return 0;
 }
 
-/* Should be called with nsid_lock held. If a new id is assigned, the bool alloc
- * is set to true, thus the caller knows that the new id must be notified via
- * rtnl.
- */
-static int __peernet2id_alloc(struct net *net, struct net *peer, bool *alloc)
+/* Must be called from RCU-critical section or with nsid_lock held */
+static int __peernet2id(const struct net *net, struct net *peer)
 {
 	int id = idr_for_each(&net->netns_ids, net_eq_idr, peer);
-	bool alloc_it = *alloc;
-
-	*alloc = false;
 
 	/* Magic value for id 0. */
 	if (id == NET_ID_ZERO)
@@ -209,61 +303,60 @@ static int __peernet2id_alloc(struct net *net, struct net *peer, bool *alloc)
 	if (id > 0)
 		return id;
 
-	if (alloc_it) {
-		id = alloc_netid(net, peer, -1);
-		*alloc = true;
-		return id >= 0 ? id : NETNSA_NSID_NOT_ASSIGNED;
-	}
-
 	return NETNSA_NSID_NOT_ASSIGNED;
 }
 
-/* should be called with nsid_lock held */
-static int __peernet2id(struct net *net, struct net *peer)
-{
-	bool no = false;
-
-	return __peernet2id_alloc(net, peer, &no);
-}
-
-static void rtnl_net_notifyid(struct net *net, int cmd, int id);
+static void rtnl_net_notifyid(struct net *net, int cmd, int id, u32 portid,
+			      struct nlmsghdr *nlh, gfp_t gfp);
 /* This function returns the id of a peer netns. If no id is assigned, one will
  * be allocated and returned.
  */
-int peernet2id_alloc(struct net *net, struct net *peer)
+int peernet2id_alloc(struct net *net, struct net *peer, gfp_t gfp)
 {
-	bool alloc = false, alive = false;
 	int id;
 
-	if (refcount_read(&net->count) == 0)
+	if (!check_net(net))
 		return NETNSA_NSID_NOT_ASSIGNED;
-	spin_lock_bh(&net->nsid_lock);
-	/*
-	 * When peer is obtained from RCU lists, we may race with
+
+	spin_lock(&net->nsid_lock);
+	id = __peernet2id(net, peer);
+	if (id >= 0) {
+		spin_unlock(&net->nsid_lock);
+		return id;
+	}
+
+	/* When peer is obtained from RCU lists, we may race with
 	 * its cleanup. Check whether it's alive, and this guarantees
 	 * we never hash a peer back to net->netns_ids, after it has
 	 * just been idr_remove()'d from there in cleanup_net().
 	 */
-	if (maybe_get_net(peer))
-		alive = alloc = true;
-	id = __peernet2id_alloc(net, peer, &alloc);
-	spin_unlock_bh(&net->nsid_lock);
-	if (alloc && id >= 0)
-		rtnl_net_notifyid(net, RTM_NEWNSID, id);
-	if (alive)
-		put_net(peer);
+	if (!maybe_get_net(peer)) {
+		spin_unlock(&net->nsid_lock);
+		return NETNSA_NSID_NOT_ASSIGNED;
+	}
+
+	id = alloc_netid(net, peer, -1);
+	spin_unlock(&net->nsid_lock);
+
+	put_net(peer);
+	if (id < 0)
+		return NETNSA_NSID_NOT_ASSIGNED;
+
+	rtnl_net_notifyid(net, RTM_NEWNSID, id, 0, NULL, gfp);
+
 	return id;
 }
 EXPORT_SYMBOL_GPL(peernet2id_alloc);
 
 /* This function returns, if assigned, the id of a peer netns. */
-int peernet2id(struct net *net, struct net *peer)
+int peernet2id(const struct net *net, struct net *peer)
 {
 	int id;
 
-	spin_lock_bh(&net->nsid_lock);
+	rcu_read_lock();
 	id = __peernet2id(net, peer);
-	spin_unlock_bh(&net->nsid_lock);
+	rcu_read_unlock();
+
 	return id;
 }
 EXPORT_SYMBOL(peernet2id);
@@ -271,12 +364,12 @@ EXPORT_SYMBOL(peernet2id);
 /* This function returns true is the peer netns has an id assigned into the
  * current netns.
  */
-bool peernet_has_id(struct net *net, struct net *peer)
+bool peernet_has_id(const struct net *net, struct net *peer)
 {
 	return peernet2id(net, peer) >= 0;
 }
 
-struct net *get_net_ns_by_id(struct net *net, int id)
+struct net *get_net_ns_by_id(const struct net *net, int id)
 {
 	struct net *peer;
 
@@ -291,25 +384,64 @@ struct net *get_net_ns_by_id(struct net *net, int id)
 
 	return peer;
 }
+EXPORT_SYMBOL_GPL(get_net_ns_by_id);
 
-/*
- * 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 void preinit_net_sysctl(struct net *net)
 {
-	/* Must be called with pernet_ops_rwsem held */
-	const struct pernet_operations *ops, *saved_ops;
-	int error = 0;
-	LIST_HEAD(net_exit_list);
+	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;
+	net->core.sysctl_txq_reselection = msecs_to_jiffies(1000);
+}
+
+/* init code that must occur even if setup_net() is not called. */
+static __net_init int preinit_net(struct net *net, struct user_namespace *user_ns)
+{
+	int ret;
+
+	ret = ns_common_init(net);
+	if (ret)
+		return ret;
 
-	refcount_set(&net->count, 1);
 	refcount_set(&net->passive, 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);
+	return 0;
+}
+
+/*
+ * setup_net runs the initializers for the network namespace object.
+ */
+static __net_init int setup_net(struct net *net)
+{
+	/* Must be called with pernet_ops_rwsem held */
+	const struct pernet_operations *ops;
+	LIST_HEAD(net_exit_list);
+	int error = 0;
+
+	net->net_cookie = ns_tree_gen_id(net);
+
 	list_for_each_entry(ops, &pernet_list, list) {
 		error = ops_init(ops, net);
 		if (error < 0)
@@ -318,6 +450,7 @@ static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
 	down_write(&net_rwsem);
 	list_add_tail_rcu(&net->list, &net_namespace_list);
 	up_write(&net_rwsem);
+	ns_tree_add_raw(net);
 out:
 	return error;
 
@@ -326,38 +459,11 @@ 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_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;
-	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)
 {
@@ -385,29 +491,64 @@ static struct net *net_alloc(void)
 	if (!net)
 		goto out_free;
 
+#ifdef CONFIG_KEYS
+	net->key_domain = kzalloc(sizeof(struct key_tag), GFP_KERNEL);
+	if (!net->key_domain)
+		goto out_free_2;
+	refcount_set(&net->key_domain->usage, 1);
+#endif
+
 	rcu_assign_pointer(net->gen, ng);
 out:
 	return net;
 
+#ifdef CONFIG_KEYS
+out_free_2:
+	kmem_cache_free(net_cachep, net);
+	net = NULL;
+#endif
 out_free:
 	kfree(ng);
 	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)
 {
-	kfree(rcu_access_pointer(net->gen));
-	kmem_cache_free(net_cachep, net);
+	if (refcount_dec_and_test(&net->passive)) {
+		kfree(rcu_access_pointer(net->gen));
+
+		/* There should not be any trackers left there. */
+		ref_tracker_dir_exit(&net->notrefcnt_tracker);
+
+		/* Wait for an extra rcu_barrier() before final free. */
+		llist_add(&net->defer_free_list, &defer_free_list);
+	}
 }
 
 void net_drop_ns(void *p)
 {
-	struct net *ns = p;
-	if (ns && refcount_dec_and_test(&ns->passive))
-		net_free(ns);
+	struct net *net = (struct net *)p;
+
+	if (net)
+		net_passive_dec(net);
 }
 
-struct net *copy_net_ns(unsigned long flags,
+struct net *copy_net_ns(u64 flags,
 			struct user_namespace *user_ns, struct net *old_net)
 {
 	struct ucounts *ucounts;
@@ -426,7 +567,10 @@ struct net *copy_net_ns(unsigned long flags,
 		rv = -ENOMEM;
 		goto dec_ucounts;
 	}
-	refcount_set(&net->passive, 1);
+
+	rv = preinit_net(net, user_ns);
+	if (rv < 0)
+		goto dec_ucounts;
 	net->ucounts = ucounts;
 	get_user_ns(user_ns);
 
@@ -434,14 +578,18 @@ 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);
 
 	if (rv < 0) {
 put_userns:
+		ns_common_free(net);
+#ifdef CONFIG_KEYS
+		key_remove_domain(net->key_domain);
+#endif
 		put_user_ns(user_ns);
-		net_drop_ns(net);
+		net_passive_dec(net);
 dec_ucounts:
 		dec_net_namespaces(ucounts);
 		return ERR_PTR(rv);
@@ -488,30 +636,34 @@ static void unhash_nsid(struct net *net, struct net *last)
 	for_each_net(tmp) {
 		int id;
 
-		spin_lock_bh(&tmp->nsid_lock);
+		spin_lock(&tmp->nsid_lock);
 		id = __peernet2id(tmp, net);
 		if (id >= 0)
 			idr_remove(&tmp->netns_ids, id);
-		spin_unlock_bh(&tmp->nsid_lock);
+		spin_unlock(&tmp->nsid_lock);
 		if (id >= 0)
-			rtnl_net_notifyid(tmp, RTM_DELNSID, id);
+			rtnl_net_notifyid(tmp, RTM_DELNSID, id, 0, NULL,
+					  GFP_KERNEL);
 		if (tmp == last)
 			break;
 	}
-	spin_lock_bh(&net->nsid_lock);
+	spin_lock(&net->nsid_lock);
 	idr_destroy(&net->netns_ids);
-	spin_unlock_bh(&net->nsid_lock);
+	spin_unlock(&net->nsid_lock);
 }
 
 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);
 
+	WRITE_ONCE(cleanup_net_task, current);
+
 	/* Atomically snapshot the list of namespaces to cleanup */
 	net_kill_list = llist_del_all(&cleanup_list);
 
@@ -519,8 +671,10 @@ static void cleanup_net(struct work_struct *work)
 
 	/* Don't let anyone else find us. */
 	down_write(&net_rwsem);
-	llist_for_each_entry(net, net_kill_list, cleanup_list)
+	llist_for_each_entry(net, net_kill_list, cleanup_list) {
+		ns_tree_remove(net);
 		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
 	 * to a net from net_kill_list (see peernet2id_alloc()).
@@ -539,20 +693,7 @@ static void cleanup_net(struct work_struct *work)
 		list_add_tail(&net->exit_list, &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.
-	 */
-	synchronize_rcu();
-
-	/* 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);
 
@@ -561,13 +702,20 @@ 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);
+		ns_common_free(net);
 		dec_net_namespaces(net->ucounts);
+#ifdef CONFIG_KEYS
+		key_remove_domain(net->key_domain);
+#endif
 		put_user_ns(net->user_ns);
-		net_drop_ns(net);
+		net_passive_dec(net);
 	}
+	WRITE_ONCE(cleanup_net_task, NULL);
 }
 
 /**
@@ -590,39 +738,47 @@ static DECLARE_WORK(net_cleanup_work, cleanup_net);
 
 void __put_net(struct net *net)
 {
+	ref_tracker_dir_exit(&net->refcnt_tracker);
 	/* Cleanup the network namespace in process context */
 	if (llist_add(&net->cleanup_list, &cleanup_list))
 		queue_work(netns_wq, &net_cleanup_work);
 }
 EXPORT_SYMBOL_GPL(__put_net);
 
-struct net *get_net_ns_by_fd(int fd)
+/**
+ * get_net_ns - increment the refcount of the network namespace
+ * @ns: common namespace (net)
+ *
+ * Returns the net's common namespace or ERR_PTR() if ref is zero.
+ */
+struct ns_common *get_net_ns(struct ns_common *ns)
 {
-	struct file *file;
-	struct ns_common *ns;
 	struct net *net;
 
-	file = proc_ns_fget(fd);
-	if (IS_ERR(file))
-		return ERR_CAST(file);
-
-	ns = get_proc_ns(file_inode(file));
-	if (ns->ops == &netns_operations)
-		net = get_net(container_of(ns, struct net, ns));
-	else
-		net = ERR_PTR(-EINVAL);
-
-	fput(file);
-	return 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);
 
-#else
 struct net *get_net_ns_by_fd(int fd)
 {
+	CLASS(fd, f)(fd);
+
+	if (fd_empty(f))
+		return ERR_PTR(-EBADF);
+
+	if (proc_ns_file(fd_file(f))) {
+		struct ns_common *ns = get_proc_ns(file_inode(fd_file(f)));
+		if (ns->ops == &netns_operations)
+			return get_net(container_of(ns, struct net, ns));
+	}
+
 	return ERR_PTR(-EINVAL);
 }
-#endif
 EXPORT_SYMBOL_GPL(get_net_ns_by_fd);
+#endif
 
 struct net *get_net_ns_by_pid(pid_t pid)
 {
@@ -646,22 +802,37 @@ struct net *get_net_ns_by_pid(pid_t pid)
 }
 EXPORT_SYMBOL_GPL(get_net_ns_by_pid);
 
-static __net_init int net_ns_net_init(struct net *net)
+#ifdef CONFIG_NET_NS_REFCNT_TRACKER
+static void net_ns_net_debugfs(struct net *net)
 {
-#ifdef CONFIG_NET_NS
-	net->ns.ops = &netns_operations;
-#endif
-	return ns_alloc_inum(&net->ns);
+	ref_tracker_dir_symlink(&net->refcnt_tracker, "netns-%llx-%u-refcnt",
+				net->net_cookie, net->ns.inum);
+	ref_tracker_dir_symlink(&net->notrefcnt_tracker, "netns-%llx-%u-notrefcnt",
+				net->net_cookie, net->ns.inum);
+}
+
+static int __init init_net_debugfs(void)
+{
+	ref_tracker_dir_debugfs(&init_net.refcnt_tracker);
+	ref_tracker_dir_debugfs(&init_net.notrefcnt_tracker);
+	net_ns_net_debugfs(&init_net);
+	return 0;
 }
+late_initcall(init_net_debugfs);
+#else
+static void net_ns_net_debugfs(struct net *net)
+{
+}
+#endif
 
-static __net_exit void net_ns_net_exit(struct net *net)
+static __net_init int net_ns_net_init(struct net *net)
 {
-	ns_free_inum(&net->ns);
+	net_ns_net_debugfs(net);
+	return 0;
 }
 
 static struct pernet_operations __net_initdata net_ns_ops = {
 	.init = net_ns_net_init,
-	.exit = net_ns_net_exit,
 };
 
 static const struct nla_policy rtnl_net_policy[NETNSA_MAX + 1] = {
@@ -681,8 +852,8 @@ static int rtnl_net_newid(struct sk_buff *skb, struct nlmsghdr *nlh,
 	struct net *peer;
 	int nsid, err;
 
-	err = nlmsg_parse(nlh, sizeof(struct rtgenmsg), tb, NETNSA_MAX,
-			  rtnl_net_policy, extack);
+	err = nlmsg_parse_deprecated(nlh, sizeof(struct rtgenmsg), tb,
+				     NETNSA_MAX, rtnl_net_policy, extack);
 	if (err < 0)
 		return err;
 	if (!tb[NETNSA_NSID]) {
@@ -707,9 +878,9 @@ static int rtnl_net_newid(struct sk_buff *skb, struct nlmsghdr *nlh,
 		return PTR_ERR(peer);
 	}
 
-	spin_lock_bh(&net->nsid_lock);
+	spin_lock(&net->nsid_lock);
 	if (__peernet2id(net, peer) >= 0) {
-		spin_unlock_bh(&net->nsid_lock);
+		spin_unlock(&net->nsid_lock);
 		err = -EEXIST;
 		NL_SET_BAD_ATTR(extack, nla);
 		NL_SET_ERR_MSG(extack,
@@ -718,9 +889,10 @@ static int rtnl_net_newid(struct sk_buff *skb, struct nlmsghdr *nlh,
 	}
 
 	err = alloc_netid(net, peer, nsid);
-	spin_unlock_bh(&net->nsid_lock);
+	spin_unlock(&net->nsid_lock);
 	if (err >= 0) {
-		rtnl_net_notifyid(net, RTM_NEWNSID, err);
+		rtnl_net_notifyid(net, RTM_NEWNSID, err, NETLINK_CB(skb).portid,
+				  nlh, GFP_KERNEL);
 		err = 0;
 	} else if (err == -ENOSPC && nsid >= 0) {
 		err = -EEXIST;
@@ -786,11 +958,13 @@ static int rtnl_net_valid_getid_req(struct sk_buff *skb,
 	int i, err;
 
 	if (!netlink_strict_get_check(skb))
-		return nlmsg_parse(nlh, sizeof(struct rtgenmsg), tb, NETNSA_MAX,
-				   rtnl_net_policy, extack);
+		return nlmsg_parse_deprecated(nlh, sizeof(struct rtgenmsg),
+					      tb, NETNSA_MAX, rtnl_net_policy,
+					      extack);
 
-	err = nlmsg_parse_strict(nlh, sizeof(struct rtgenmsg), tb, NETNSA_MAX,
-				 rtnl_net_policy, extack);
+	err = nlmsg_parse_deprecated_strict(nlh, sizeof(struct rtgenmsg), tb,
+					    NETNSA_MAX, rtnl_net_policy,
+					    extack);
 	if (err)
 		return err;
 
@@ -838,7 +1012,7 @@ static int rtnl_net_getid(struct sk_buff *skb, struct nlmsghdr *nlh,
 		peer = get_net_ns_by_fd(nla_get_u32(tb[NETNSA_FD]));
 		nla = tb[NETNSA_FD];
 	} else if (tb[NETNSA_NSID]) {
-		peer = get_net_ns_by_id(net, nla_get_u32(tb[NETNSA_NSID]));
+		peer = get_net_ns_by_id(net, nla_get_s32(tb[NETNSA_NSID]));
 		if (!peer)
 			peer = ERR_PTR(-ENOENT);
 		nla = tb[NETNSA_NSID];
@@ -900,6 +1074,7 @@ struct rtnl_net_dump_cb {
 	int s_idx;
 };
 
+/* Runs in RCU-critical section. */
 static int rtnl_net_dumpid_one(int id, void *peer, void *data)
 {
 	struct rtnl_net_dump_cb *net_cb = (struct rtnl_net_dump_cb *)data;
@@ -928,8 +1103,9 @@ static int rtnl_valid_dump_net_req(const struct nlmsghdr *nlh, struct sock *sk,
 	struct nlattr *tb[NETNSA_MAX + 1];
 	int err, i;
 
-	err = nlmsg_parse_strict(nlh, sizeof(struct rtgenmsg), tb, NETNSA_MAX,
-				 rtnl_net_policy, extack);
+	err = nlmsg_parse_deprecated_strict(nlh, sizeof(struct rtgenmsg), tb,
+					    NETNSA_MAX, rtnl_net_policy,
+					    extack);
 	if (err < 0)
 		return err;
 
@@ -983,37 +1159,30 @@ static int rtnl_net_dumpid(struct sk_buff *skb, struct netlink_callback *cb)
 			goto end;
 	}
 
-	spin_lock_bh(&net_cb.tgt_net->nsid_lock);
-	if (net_cb.fillargs.add_ref &&
-	    !net_eq(net_cb.ref_net, net_cb.tgt_net) &&
-	    !spin_trylock_bh(&net_cb.ref_net->nsid_lock)) {
-		spin_unlock_bh(&net_cb.tgt_net->nsid_lock);
-		err = -EAGAIN;
-		goto end;
-	}
+	rcu_read_lock();
 	idr_for_each(&net_cb.tgt_net->netns_ids, rtnl_net_dumpid_one, &net_cb);
-	if (net_cb.fillargs.add_ref &&
-	    !net_eq(net_cb.ref_net, net_cb.tgt_net))
-		spin_unlock_bh(&net_cb.ref_net->nsid_lock);
-	spin_unlock_bh(&net_cb.tgt_net->nsid_lock);
+	rcu_read_unlock();
 
 	cb->args[0] = net_cb.idx;
 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)
+static void rtnl_net_notifyid(struct net *net, int cmd, int id, u32 portid,
+			      struct nlmsghdr *nlh, gfp_t gfp)
 {
 	struct net_fill_args fillargs = {
+		.portid = portid,
+		.seq = nlh ? nlh->nlmsg_seq : 0,
 		.cmd = cmd,
 		.nsid = id,
 	};
 	struct sk_buff *msg;
 	int err = -ENOMEM;
 
-	msg = nlmsg_new(rtnl_net_get_size(), GFP_KERNEL);
+	msg = nlmsg_new(rtnl_net_get_size(), gfp);
 	if (!msg)
 		goto out;
 
@@ -1021,7 +1190,7 @@ static void rtnl_net_notifyid(struct net *net, int cmd, int id)
 	if (err < 0)
 		goto err_out;
 
-	rtnl_notify(msg, net, 0, RTNLGRP_NSID, NULL, 0);
+	rtnl_notify(msg, net, portid, RTNLGRP_NSID, nlh, gfp);
 	return;
 
 err_out:
@@ -1030,11 +1199,63 @@ out:
 	rtnl_set_sk_err(net, RTNLGRP_NSID, err);
 }
 
-static int __init net_ns_init(void)
+#ifdef CONFIG_NET_NS
+static void __init netns_ipv4_struct_check(void)
+{
+	/* TX readonly hotpath cache lines */
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_early_retrans);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_tso_win_divisor);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_tso_rtt_log);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_autocorking);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_min_snd_mss);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_notsent_lowat);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_limit_output_bytes);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_min_rtt_wlen);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_tcp_wmem);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_tx,
+				      sysctl_ip_fwd_use_pmtu);
+
+	/* RX readonly hotpath cache line */
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
+				      sysctl_tcp_moderate_rcvbuf);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
+				      sysctl_tcp_rcvbuf_low_rtt);
+	CACHELINE_ASSERT_GROUP_MEMBER(struct netns_ipv4, netns_ipv4_read_rx,
+				      sysctl_ip_early_demux);
+	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);
+}
+#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;
 
 #ifdef CONFIG_NET_NS
+	netns_ipv4_struct_check();
 	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
 					SMP_CACHE_BYTES,
 					SLAB_PANIC|SLAB_ACCOUNT, NULL);
@@ -1051,8 +1272,18 @@ static int __init net_ns_init(void)
 
 	rcu_assign_pointer(init_net.gen, ng);
 
+#ifdef CONFIG_KEYS
+	init_net.key_domain = &init_net_key_domain;
+#endif
+	/*
+	 * This currently cannot fail as the initial network namespace
+	 * has a static inode number.
+	 */
+	if (preinit_net(&init_net, &init_user_ns))
+		panic("Could not preinitialize the initial network namespace");
+
 	down_write(&pernet_ops_rwsem);
-	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;
@@ -1061,26 +1292,19 @@ static int __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);
-
-	return 0;
+	rtnl_register_many(net_ns_rtnl_msg_handlers);
 }
 
-pure_initcall(net_ns_init);
-
 #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.
 		 */
@@ -1096,22 +1320,21 @@ 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);
-	ops_exit_list(ops, &net_exit_list);
-	ops_free_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);
-	ops_exit_list(ops, &net_exit_list);
-	ops_free_list(ops, &net_exit_list);
+
+	list_del(&ops->list);
+	ops_undo_single(ops, &net_exit_list);
 }
 
 #else
@@ -1133,9 +1356,9 @@ 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);
-		ops_exit_list(ops, &net_exit_list);
-		ops_free_list(ops, &net_exit_list);
+		ops_undo_single(ops, &net_exit_list);
 	}
 }
 
@@ -1148,13 +1371,20 @@ 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) {
@@ -1285,22 +1515,18 @@ static struct ns_common *netns_get(struct task_struct *task)
 	return net ? &net->ns : NULL;
 }
 
-static inline struct net *to_net_ns(struct ns_common *ns)
-{
-	return container_of(ns, struct net, ns);
-}
-
 static void netns_put(struct ns_common *ns)
 {
 	put_net(to_net_ns(ns));
 }
 
-static int netns_install(struct nsproxy *nsproxy, struct ns_common *ns)
+static int netns_install(struct nsset *nsset, struct ns_common *ns)
 {
+	struct nsproxy *nsproxy = nsset->nsproxy;
 	struct net *net = to_net_ns(ns);
 
 	if (!ns_capable(net->user_ns, CAP_SYS_ADMIN) ||
-	    !ns_capable(current_user_ns(), CAP_SYS_ADMIN))
+	    !ns_capable(nsset->cred->user_ns, CAP_SYS_ADMIN))
 		return -EPERM;
 
 	put_net(nsproxy->net_ns);
@@ -1315,7 +1541,6 @@ static struct user_namespace *netns_owner(struct ns_common *ns)
 
 const struct proc_ns_operations netns_operations = {
 	.name		= "net",
-	.type		= CLONE_NEWNET,
 	.get		= netns_get,
 	.put		= netns_put,
 	.install	= netns_install,