tcp/dccp: fix possible race __inet_lookup_established()

Michal Kubecek and Firo Yang did a very nice analysis of crashes
happening in __inet_lookup_established().

Since a TCP socket can go from TCP_ESTABLISH to TCP_LISTEN
(via a close()/socket()/listen() cycle) without a RCU grace period,
I should not have changed listeners linkage in their hash table.

They must use the nulls protocol (Documentation/RCU/rculist_nulls.txt),
so that a lookup can detect a socket in a hash list was moved in
another one.

Since we added code in commit d296ba60d8e2 ("soreuseport: Resolve
merge conflict for v4/v6 ordering fix"), we have to add
hlist_nulls_add_tail_rcu() helper.

Fixes: 3b24d854cb35 ("tcp/dccp: do not touch listener sk_refcnt under synflood")
Signed-off-by: Eric Dumazet <[e-mail]>
Reported-by: Michal Kubecek <[e-mail]>
Reported-by: Firo Yang <[e-mail]>
Reviewed-by: Michal Kubecek <[e-mail]>
Link: https://lore.kernel.org/netdev/[e-mail]/
Signed-off-by: Jakub Kicinski <[e-mail]>

include/linux/rculist_nulls.h

···
 }
 
 /**
+ * hlist_nulls_add_tail_rcu
+ * @n: the element to add to the hash list.
+ * @h: the list to add to.
+ *
+ * Description:
+ * Adds the specified element to the specified hlist_nulls,
+ * while permitting racing traversals.
+ *
+ * The caller must take whatever precautions are necessary
+ * (such as holding appropriate locks) to avoid racing
+ * with another list-mutation primitive, such as hlist_nulls_add_head_rcu()
+ * or hlist_nulls_del_rcu(), running on this same list.
+ * However, it is perfectly legal to run concurrently with
+ * the _rcu list-traversal primitives, such as
+ * hlist_nulls_for_each_entry_rcu(), used to prevent memory-consistency
+ * problems on Alpha CPUs.  Regardless of the type of CPU, the
+ * list-traversal primitive must be guarded by rcu_read_lock().
+ */
+static inline void hlist_nulls_add_tail_rcu(struct hlist_nulls_node *n,
+					    struct hlist_nulls_head *h)
+{
+	struct hlist_nulls_node *i, *last = NULL;
+
+	/* Note: write side code, so rcu accessors are not needed. */
+	for (i = h->first; !is_a_nulls(i); i = i->next)
+		last = i;
+
+	if (last) {
+		n->next = last->next;
+		n->pprev = &last->next;
+		rcu_assign_pointer(hlist_next_rcu(last), n);
+	} else {
+		hlist_nulls_add_head_rcu(n, h);
+	}
+}
+
+/**
  * hlist_nulls_for_each_entry_rcu - iterate over rcu list of given type
  * @tpos:	the type * to use as a loop cursor.
  * @pos:	the &struct hlist_nulls_node to use as a loop cursor.

include/net/inet_hashtables.h

···
 	struct hlist_head	chain;
 };
 
-/*
- * Sockets can be hashed in established or listening table
+/* Sockets can be hashed in established or listening table.
+ * We must use different 'nulls' end-of-chain value for all hash buckets :
+ * A socket might transition from ESTABLISH to LISTEN state without
+ * RCU grace period. A lookup in ehash table needs to handle this case.
  */
+#define LISTENING_NULLS_BASE (1U << 29)
 struct inet_listen_hashbucket {
 	spinlock_t		lock;
 	unsigned int		count;
-	struct hlist_head	head;
+	union {
+		struct hlist_head	head;
+		struct hlist_nulls_head	nulls_head;
+	};
 };
 
 /* This is for listening sockets, thus all sockets which possess wildcards. */

include/net/sock.h

···
 	hlist_nulls_add_head_rcu(&sk->sk_nulls_node, list);
 }
 
+static inline void __sk_nulls_add_node_tail_rcu(struct sock *sk, struct hlist_nulls_head *list)
+{
+	hlist_nulls_add_tail_rcu(&sk->sk_nulls_node, list);
+}
+
 static inline void sk_nulls_add_node_rcu(struct sock *sk, struct hlist_nulls_head *list)
 {
 	sock_hold(sk);

net/ipv4/inet_diag.c

···
 
 		for (i = s_i; i < INET_LHTABLE_SIZE; i++) {
 			struct inet_listen_hashbucket *ilb;
+			struct hlist_nulls_node *node;
 
 			num = 0;
 			ilb = &hashinfo->listening_hash[i];
 			spin_lock(&ilb->lock);
-			sk_for_each(sk, &ilb->head) {
+			sk_nulls_for_each(sk, node, &ilb->nulls_head) {
 				struct inet_sock *inet = inet_sk(sk);
 
 				if (!net_eq(sock_net(sk), net))

net/ipv4/inet_hashtables.c

···
 				   struct inet_listen_hashbucket *ilb)
 {
 	struct inet_bind_bucket *tb = inet_csk(sk)->icsk_bind_hash;
+	const struct hlist_nulls_node *node;
 	struct sock *sk2;
 	kuid_t uid = sock_i_uid(sk);
 
-	sk_for_each_rcu(sk2, &ilb->head) {
+	sk_nulls_for_each_rcu(sk2, node, &ilb->nulls_head) {
 		if (sk2 != sk &&
 		    sk2->sk_family == sk->sk_family &&
 		    ipv6_only_sock(sk2) == ipv6_only_sock(sk) &&
···
 	}
 	if (IS_ENABLED(CONFIG_IPV6) && sk->sk_reuseport &&
 		sk->sk_family == AF_INET6)
-		hlist_add_tail_rcu(&sk->sk_node, &ilb->head);
+		__sk_nulls_add_node_tail_rcu(sk, &ilb->nulls_head);
 	else
-		hlist_add_head_rcu(&sk->sk_node, &ilb->head);
+		__sk_nulls_add_node_rcu(sk, &ilb->nulls_head);
 	inet_hash2(hashinfo, sk);
 	ilb->count++;
 	sock_set_flag(sk, SOCK_RCU_FREE);
···
 		reuseport_detach_sock(sk);
 	if (ilb) {
 		inet_unhash2(hashinfo, sk);
-		 __sk_del_node_init(sk);
-		 ilb->count--;
-	} else {
-		__sk_nulls_del_node_init_rcu(sk);
+		ilb->count--;
 	}
+	__sk_nulls_del_node_init_rcu(sk);
 	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
 unlock:
 	spin_unlock_bh(lock);
···
 
 	for (i = 0; i < INET_LHTABLE_SIZE; i++) {
 		spin_lock_init(&h->listening_hash[i].lock);
-		INIT_HLIST_HEAD(&h->listening_hash[i].head);
+		INIT_HLIST_NULLS_HEAD(&h->listening_hash[i].nulls_head,
+				      i + LISTENING_NULLS_BASE);
 		h->listening_hash[i].count = 0;
 	}
 

net/ipv4/tcp_ipv4.c

···
 	struct tcp_iter_state *st = seq->private;
 	struct net *net = seq_file_net(seq);
 	struct inet_listen_hashbucket *ilb;
+	struct hlist_nulls_node *node;
 	struct sock *sk = cur;
 
 	if (!sk) {
 get_head:
 		ilb = &tcp_hashinfo.listening_hash[st->bucket];
 		spin_lock(&ilb->lock);
-		sk = sk_head(&ilb->head);
+		sk = sk_nulls_head(&ilb->nulls_head);
 		st->offset = 0;
 		goto get_sk;
 	}
···
 	++st->num;
 	++st->offset;
 
-	sk = sk_next(sk);
+	sk = sk_nulls_next(sk);
 get_sk:
-	sk_for_each_from(sk) {
+	sk_nulls_for_each_from(sk, node) {
 		if (!net_eq(sock_net(sk), net))
 			continue;
 		if (sk->sk_family == afinfo->family)