7 files changed, 85 insertions, 77 deletions

diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index e3b4aee4244e..2c7c299ee2b9 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -830,7 +830,7 @@ static struct key_vector *resize(struct trie *t, struct key_vector *tn)
 	/* Double as long as the resulting node has a number of
 	 * nonempty nodes that are above the threshold.
 	 */
-	while (should_inflate(tp, tn) && max_work--) {
+	while (should_inflate(tp, tn) && max_work) {
 		tp = inflate(t, tn);
 		if (!tp) {
 #ifdef CONFIG_IP_FIB_TRIE_STATS
@@ -839,17 +839,21 @@ static struct key_vector *resize(struct trie *t, struct key_vector *tn)
 			break;
 		}
 
+		max_work--;
 		tn = get_child(tp, cindex);
 	}
 
+	/* update parent in case inflate failed */
+	tp = node_parent(tn);
+
 	/* Return if at least one inflate is run */
 	if (max_work != MAX_WORK)
-		return node_parent(tn);
+		return tp;
 
 	/* Halve as long as the number of empty children in this
 	 * node is above threshold.
 	 */
-	while (should_halve(tp, tn) && max_work--) {
+	while (should_halve(tp, tn) && max_work) {
 		tp = halve(t, tn);
 		if (!tp) {
 #ifdef CONFIG_IP_FIB_TRIE_STATS
@@ -858,6 +862,7 @@ static struct key_vector *resize(struct trie *t, struct key_vector *tn)
 			break;
 		}
 
+		max_work--;
 		tn = get_child(tp, cindex);
 	}
 
@@ -865,7 +870,7 @@ static struct key_vector *resize(struct trie *t, struct key_vector *tn)
 	if (should_collapse(tn))
 		return collapse(t, tn);
 
-	/* update parent in case inflate or halve failed */
+	/* update parent in case halve failed */
 	tp = node_parent(tn);
 
 	/* Return if at least one deflate was run */
diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 844808d9337b..79c0c9439fdc 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -403,18 +403,17 @@ struct dst_entry *inet_csk_route_req(struct sock *sk,
 				     struct flowi4 *fl4,
 				     const struct request_sock *req)
 {
-	struct rtable *rt;
 	const struct inet_request_sock *ireq = inet_rsk(req);
-	struct ip_options_rcu *opt = inet_rsk(req)->opt;
-	struct net *net = sock_net(sk);
-	int flags = inet_sk_flowi_flags(sk);
+	struct net *net = read_pnet(&ireq->ireq_net);
+	struct ip_options_rcu *opt = ireq->opt;
+	struct rtable *rt;
 
-	flowi4_init_output(fl4, sk->sk_bound_dev_if, ireq->ir_mark,
+	flowi4_init_output(fl4, ireq->ir_iif, ireq->ir_mark,
 			   RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
-			   sk->sk_protocol,
-			   flags,
+			   sk->sk_protocol, inet_sk_flowi_flags(sk),
 			   (opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
-			   ireq->ir_loc_addr, ireq->ir_rmt_port, inet_sk(sk)->inet_sport);
+			   ireq->ir_loc_addr, ireq->ir_rmt_port,
+			   htons(ireq->ir_num));
 	security_req_classify_flow(req, flowi4_to_flowi(fl4));
 	rt = ip_route_output_flow(net, fl4, sk);
 	if (IS_ERR(rt))
@@ -436,9 +435,9 @@ struct dst_entry *inet_csk_route_child_sock(struct sock *sk,
 					    const struct request_sock *req)
 {
 	const struct inet_request_sock *ireq = inet_rsk(req);
+	struct net *net = read_pnet(&ireq->ireq_net);
 	struct inet_sock *newinet = inet_sk(newsk);
 	struct ip_options_rcu *opt;
-	struct net *net = sock_net(sk);
 	struct flowi4 *fl4;
 	struct rtable *rt;
 
@@ -446,11 +445,12 @@ struct dst_entry *inet_csk_route_child_sock(struct sock *sk,
 
 	rcu_read_lock();
 	opt = rcu_dereference(newinet->inet_opt);
-	flowi4_init_output(fl4, sk->sk_bound_dev_if, inet_rsk(req)->ir_mark,
+	flowi4_init_output(fl4, ireq->ir_iif, ireq->ir_mark,
 			   RT_CONN_FLAGS(sk), RT_SCOPE_UNIVERSE,
 			   sk->sk_protocol, inet_sk_flowi_flags(sk),
 			   (opt && opt->opt.srr) ? opt->opt.faddr : ireq->ir_rmt_addr,
-			   ireq->ir_loc_addr, ireq->ir_rmt_port, inet_sk(sk)->inet_sport);
+			   ireq->ir_loc_addr, ireq->ir_rmt_port,
+			   htons(ireq->ir_num));
 	security_req_classify_flow(req, flowi4_to_flowi(fl4));
 	rt = ip_route_output_flow(net, fl4, sk);
 	if (IS_ERR(rt))
@@ -495,7 +495,7 @@ struct request_sock *inet_csk_search_req(struct sock *sk,
 	u32 hash = inet_synq_hash(raddr, rport, lopt->hash_rnd,
 				  lopt->nr_table_entries);
 
-	write_lock(&icsk->icsk_accept_queue.syn_wait_lock);
+	spin_lock(&icsk->icsk_accept_queue.syn_wait_lock);
 	for (req = lopt->syn_table[hash]; req != NULL; req = req->dl_next) {
 		const struct inet_request_sock *ireq = inet_rsk(req);
 
@@ -508,7 +508,7 @@ struct request_sock *inet_csk_search_req(struct sock *sk,
 			break;
 		}
 	}
-	write_unlock(&icsk->icsk_accept_queue.syn_wait_lock);
+	spin_unlock(&icsk->icsk_accept_queue.syn_wait_lock);
 
 	return req;
 }
@@ -571,8 +571,9 @@ static void reqsk_timer_handler(unsigned long data)
 	struct inet_connection_sock *icsk = inet_csk(sk_listener);
 	struct request_sock_queue *queue = &icsk->icsk_accept_queue;
 	struct listen_sock *lopt = queue->listen_opt;
-	int expire = 0, resend = 0;
+	int qlen, expire = 0, resend = 0;
 	int max_retries, thresh;
+	u8 defer_accept;
 
 	if (sk_listener->sk_state != TCP_LISTEN || !lopt) {
 		reqsk_put(req);
@@ -598,21 +599,23 @@ static void reqsk_timer_handler(unsigned long data)
 	 * embrions; and abort old ones without pity, if old
 	 * ones are about to clog our table.
 	 */
-	if (listen_sock_qlen(lopt) >> (lopt->max_qlen_log - 1)) {
+	qlen = listen_sock_qlen(lopt);
+	if (qlen >> (lopt->max_qlen_log - 1)) {
 		int young = listen_sock_young(lopt) << 1;
 
 		while (thresh > 2) {
-			if (listen_sock_qlen(lopt) < young)
+			if (qlen < young)
 				break;
 			thresh--;
 			young <<= 1;
 		}
 	}
-	if (queue->rskq_defer_accept)
-		max_retries = queue->rskq_defer_accept;
-	syn_ack_recalc(req, thresh, max_retries, queue->rskq_defer_accept,
+	defer_accept = READ_ONCE(queue->rskq_defer_accept);
+	if (defer_accept)
+		max_retries = defer_accept;
+	syn_ack_recalc(req, thresh, max_retries, defer_accept,
 		       &expire, &resend);
-	req->rsk_ops->syn_ack_timeout(sk_listener, req);
+	req->rsk_ops->syn_ack_timeout(req);
 	if (!expire &&
 	    (!resend ||
 	     !inet_rtx_syn_ack(sk_listener, req) ||
@@ -647,10 +650,10 @@ void reqsk_queue_hash_req(struct request_sock_queue *queue,
 	setup_timer(&req->rsk_timer, reqsk_timer_handler, (unsigned long)req);
 	req->rsk_hash = hash;
 
-	write_lock(&queue->syn_wait_lock);
+	spin_lock(&queue->syn_wait_lock);
 	req->dl_next = lopt->syn_table[hash];
 	lopt->syn_table[hash] = req;
-	write_unlock(&queue->syn_wait_lock);
+	spin_unlock(&queue->syn_wait_lock);
 
 	mod_timer_pinned(&req->rsk_timer, jiffies + timeout);
 }
diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c
index f984b2001d0a..76322c9867d5 100644
--- a/net/ipv4/inet_diag.c
+++ b/net/ipv4/inet_diag.c
@@ -728,7 +728,7 @@ static int inet_diag_dump_reqs(struct sk_buff *skb, struct sock *sk,
 
 	entry.family = sk->sk_family;
 
-	read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+	spin_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 
 	lopt = icsk->icsk_accept_queue.listen_opt;
 	if (!lopt || !listen_sock_qlen(lopt))
@@ -776,7 +776,7 @@ static int inet_diag_dump_reqs(struct sk_buff *skb, struct sock *sk,
 	}
 
 out:
-	read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+	spin_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 
 	return err;
 }
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index a7aea2048a0d..90b49e88e84a 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -636,10 +636,7 @@ slow_path:
 	left = skb->len - hlen;		/* Space per frame */
 	ptr = hlen;		/* Where to start from */
 
-	/* for bridged IP traffic encapsulated inside f.e. a vlan header,
-	 * we need to make room for the encapsulating header
-	 */
-	ll_rs = LL_RESERVED_SPACE_EXTRA(rt->dst.dev, nf_bridge_pad(skb));
+	ll_rs = LL_RESERVED_SPACE(rt->dst.dev);
 
 	/*
 	 *	Fragment the datagram.
diff --git a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
index a460a87e14f8..f0dfe92a00d6 100644
--- a/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
+++ b/net/ipv4/netfilter/nf_conntrack_l3proto_ipv4_compat.c
@@ -300,7 +300,9 @@ static int exp_seq_show(struct seq_file *s, void *v)
 		    __nf_ct_l3proto_find(exp->tuple.src.l3num),
 		    __nf_ct_l4proto_find(exp->tuple.src.l3num,
 					 exp->tuple.dst.protonum));
-	return seq_putc(s, '\n');
+	seq_putc(s, '\n');
+
+	return 0;
 }
 
 static const struct seq_operations exp_seq_ops = {
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 5554b8f33d41..4e90217003e8 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -310,6 +310,34 @@ static void do_redirect(struct sk_buff *skb, struct sock *sk)
 		dst->ops->redirect(dst, sk, skb);
 }
 
+
+/* handle ICMP messages on TCP_NEW_SYN_RECV request sockets */
+void tcp_req_err(struct sock *sk, u32 seq)
+{
+	struct request_sock *req = inet_reqsk(sk);
+	struct net *net = sock_net(sk);
+
+	/* ICMPs are not backlogged, hence we cannot get
+	 * an established socket here.
+	 */
+	WARN_ON(req->sk);
+
+	if (seq != tcp_rsk(req)->snt_isn) {
+		NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
+		reqsk_put(req);
+	} else {
+		/*
+		 * Still in SYN_RECV, just remove it silently.
+		 * There is no good way to pass the error to the newly
+		 * created socket, and POSIX does not want network
+		 * errors returned from accept().
+		 */
+		NET_INC_STATS_BH(net, LINUX_MIB_LISTENDROPS);
+		inet_csk_reqsk_queue_drop(req->rsk_listener, req);
+	}
+}
+EXPORT_SYMBOL(tcp_req_err);
+
 /*
  * This routine is called by the ICMP module when it gets some
  * sort of error condition.  If err < 0 then the socket should
@@ -343,8 +371,9 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
 	int err;
 	struct net *net = dev_net(icmp_skb->dev);
 
-	sk = inet_lookup(net, &tcp_hashinfo, iph->daddr, th->dest,
-			iph->saddr, th->source, inet_iif(icmp_skb));
+	sk = __inet_lookup_established(net, &tcp_hashinfo, iph->daddr,
+				       th->dest, iph->saddr, ntohs(th->source),
+				       inet_iif(icmp_skb));
 	if (!sk) {
 		ICMP_INC_STATS_BH(net, ICMP_MIB_INERRORS);
 		return;
@@ -353,6 +382,9 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
 		inet_twsk_put(inet_twsk(sk));
 		return;
 	}
+	seq = ntohl(th->seq);
+	if (sk->sk_state == TCP_NEW_SYN_RECV)
+		return tcp_req_err(sk, seq);
 
 	bh_lock_sock(sk);
 	/* If too many ICMPs get dropped on busy
@@ -374,7 +406,6 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
 
 	icsk = inet_csk(sk);
 	tp = tcp_sk(sk);
-	seq = ntohl(th->seq);
 	/* XXX (TFO) - tp->snd_una should be ISN (tcp_create_openreq_child() */
 	fastopen = tp->fastopen_rsk;
 	snd_una = fastopen ? tcp_rsk(fastopen)->snt_isn : tp->snd_una;
@@ -458,38 +489,6 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
 	}
 
 	switch (sk->sk_state) {
-		struct request_sock *req;
-	case TCP_LISTEN:
-		if (sock_owned_by_user(sk))
-			goto out;
-
-		req = inet_csk_search_req(sk, th->dest,
-					  iph->daddr, iph->saddr);
-		if (!req)
-			goto out;
-
-		/* ICMPs are not backlogged, hence we cannot get
-		   an established socket here.
-		 */
-		WARN_ON(req->sk);
-
-		if (seq != tcp_rsk(req)->snt_isn) {
-			NET_INC_STATS_BH(net, LINUX_MIB_OUTOFWINDOWICMPS);
-			reqsk_put(req);
-			goto out;
-		}
-
-		/*
-		 * Still in SYN_RECV, just remove it silently.
-		 * There is no good way to pass the error to the newly
-		 * created socket, and POSIX does not want network
-		 * errors returned from accept().
-		 */
-		inet_csk_reqsk_queue_drop(sk, req);
-		NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_LISTENDROPS);
-		reqsk_put(req);
-		goto out;
-
 	case TCP_SYN_SENT:
 	case TCP_SYN_RECV:
 		/* Only in fast or simultaneous open. If a fast open socket is
@@ -1909,13 +1908,13 @@ get_req:
 		}
 		sk	  = sk_nulls_next(st->syn_wait_sk);
 		st->state = TCP_SEQ_STATE_LISTENING;
-		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+		spin_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 	} else {
 		icsk = inet_csk(sk);
-		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+		spin_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 		if (reqsk_queue_len(&icsk->icsk_accept_queue))
 			goto start_req;
-		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+		spin_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 		sk = sk_nulls_next(sk);
 	}
 get_sk:
@@ -1927,7 +1926,7 @@ get_sk:
 			goto out;
 		}
 		icsk = inet_csk(sk);
-		read_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+		spin_lock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 		if (reqsk_queue_len(&icsk->icsk_accept_queue)) {
 start_req:
 			st->uid		= sock_i_uid(sk);
@@ -1936,7 +1935,7 @@ start_req:
 			st->sbucket	= 0;
 			goto get_req;
 		}
-		read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+		spin_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 	}
 	spin_unlock_bh(&ilb->lock);
 	st->offset = 0;
@@ -2155,7 +2154,7 @@ static void tcp_seq_stop(struct seq_file *seq, void *v)
 	case TCP_SEQ_STATE_OPENREQ:
 		if (v) {
 			struct inet_connection_sock *icsk = inet_csk(st->syn_wait_sk);
-			read_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
+			spin_unlock_bh(&icsk->icsk_accept_queue.syn_wait_lock);
 		}
 	case TCP_SEQ_STATE_LISTENING:
 		if (v != SEQ_START_TOKEN)
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 3daa6b5d766d..2568fd282873 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -327,7 +327,7 @@ static void tcp_fastopen_synack_timer(struct sock *sk)
 	struct request_sock *req;
 
 	req = tcp_sk(sk)->fastopen_rsk;
-	req->rsk_ops->syn_ack_timeout(sk, req);
+	req->rsk_ops->syn_ack_timeout(req);
 
 	if (req->num_timeout >= max_retries) {
 		tcp_write_err(sk);
@@ -539,9 +539,11 @@ static void tcp_write_timer(unsigned long data)
 	sock_put(sk);
 }
 
-void tcp_syn_ack_timeout(struct sock *sk, struct request_sock *req)
+void tcp_syn_ack_timeout(const struct request_sock *req)
 {
-	NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPTIMEOUTS);
+	struct net *net = read_pnet(&inet_rsk(req)->ireq_net);
+
+	NET_INC_STATS_BH(net, LINUX_MIB_TCPTIMEOUTS);
 }
 EXPORT_SYMBOL(tcp_syn_ack_timeout);