Merge branch 'listener-sock-const'

Eric Dumazet says:

====================
dccp/tcp: constify listener sock

Another patch bomb to prepare lockless TCP/DCCP LISTEN handling.

SYNACK retransmits are built and sent without listener socket
being locked. Soon, initial SYNACK packets will have same property.

This series makes sure we did not something wrong with this model,
by adding a const qualifier in all the paths taken from synack building
and transmit, for IPv4/IPv6 and TCP/dccp.

The only potential problem was the rewrite of ecn bits for connections
with DCTCP as congestion module, but this was a very minor one.
====================

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

7 files changed, 71 insertions, 51 deletions

diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 134957159c27..bac205136e1c 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -408,7 +408,7 @@ void inet_csk_reset_keepalive_timer(struct sock *sk, unsigned long len)
 }
 EXPORT_SYMBOL(inet_csk_reset_keepalive_timer);
 
-struct dst_entry *inet_csk_route_req(struct sock *sk,
+struct dst_entry *inet_csk_route_req(const struct sock *sk,
 				     struct flowi4 *fl4,
 				     const struct request_sock *req)
 {
@@ -563,7 +563,7 @@ static inline void syn_ack_recalc(struct request_sock *req, const int thresh,
 		  req->num_timeout >= rskq_defer_accept - 1;
 }
 
-int inet_rtx_syn_ack(struct sock *parent, struct request_sock *req)
+int inet_rtx_syn_ack(const struct sock *parent, struct request_sock *req)
 {
 	int err = req->rsk_ops->rtx_syn_ack(parent, req);
 
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 09a6b7bb7ea3..06d2c87ed505 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -137,7 +137,7 @@ static inline int ip_select_ttl(struct inet_sock *inet, struct dst_entry *dst)
  *		Add an ip header to a skbuff and send it out.
  *
  */
-int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
+int ip_build_and_send_pkt(struct sk_buff *skb, const struct sock *sk,
 			  __be32 saddr, __be32 daddr, struct ip_options_rcu *opt)
 {
 	struct inet_sock *inet = inet_sk(sk);
@@ -151,15 +151,17 @@ int ip_build_and_send_pkt(struct sk_buff *skb, struct sock *sk,
 	iph->version  = 4;
 	iph->ihl      = 5;
 	iph->tos      = inet->tos;
-	if (ip_dont_fragment(sk, &rt->dst))
-		iph->frag_off = htons(IP_DF);
-	else
-		iph->frag_off = 0;
 	iph->ttl      = ip_select_ttl(inet, &rt->dst);
 	iph->daddr    = (opt && opt->opt.srr ? opt->opt.faddr : daddr);
 	iph->saddr    = saddr;
 	iph->protocol = sk->sk_protocol;
-	ip_select_ident(sock_net(sk), skb, sk);
+	if (ip_dont_fragment(sk, &rt->dst)) {
+		iph->frag_off = htons(IP_DF);
+		iph->id = 0;
+	} else {
+		iph->frag_off = 0;
+		__ip_select_ident(sock_net(sk), iph, 1);
+	}
 
 	if (opt && opt->opt.optlen) {
 		iph->ihl += opt->opt.optlen>>2;
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 80f7c5b7b832..f7afcba8b1a1 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2291,7 +2291,7 @@ struct dst_entry *ipv4_blackhole_route(struct net *net, struct dst_entry *dst_or
 }
 
 struct rtable *ip_route_output_flow(struct net *net, struct flowi4 *flp4,
-				    struct sock *sk)
+				    const struct sock *sk)
 {
 	struct rtable *rt = __ip_route_output_key(net, flp4);
 
diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c
index 93c4dc3ab23f..882caa4e72bc 100644
--- a/net/ipv4/tcp_cong.c
+++ b/net/ipv4/tcp_cong.c
@@ -173,6 +173,10 @@ out:
 	 */
 	if (ca->get_info)
 		memset(icsk->icsk_ca_priv, 0, sizeof(icsk->icsk_ca_priv));
+	if (ca->flags & TCP_CONG_NEEDS_ECN)
+		INET_ECN_xmit(sk);
+	else
+		INET_ECN_dontxmit(sk);
 }
 
 void tcp_init_congestion_control(struct sock *sk)
@@ -181,6 +185,10 @@ void tcp_init_congestion_control(struct sock *sk)
 
 	if (icsk->icsk_ca_ops->init)
 		icsk->icsk_ca_ops->init(sk);
+	if (tcp_ca_needs_ecn(sk))
+		INET_ECN_xmit(sk);
+	else
+		INET_ECN_dontxmit(sk);
 }
 
 static void tcp_reinit_congestion_control(struct sock *sk,
@@ -192,8 +200,8 @@ static void tcp_reinit_congestion_control(struct sock *sk,
 	icsk->icsk_ca_ops = ca;
 	icsk->icsk_ca_setsockopt = 1;
 
-	if (sk->sk_state != TCP_CLOSE && icsk->icsk_ca_ops->init)
-		icsk->icsk_ca_ops->init(sk);
+	if (sk->sk_state != TCP_CLOSE)
+		tcp_init_congestion_control(sk);
 }
 
 /* Manage refcounts on socket close. */
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 7e2646542312..a23ba7daecbf 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -818,7 +818,7 @@ static void tcp_v4_reqsk_send_ack(struct sock *sk, struct sk_buff *skb,
  *	This still operates on a request_sock only, not on a big
  *	socket.
  */
-static int tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst,
+static int tcp_v4_send_synack(const struct sock *sk, struct dst_entry *dst,
 			      struct flowi *fl,
 			      struct request_sock *req,
 			      u16 queue_mapping,
@@ -865,7 +865,7 @@ static void tcp_v4_reqsk_destructor(struct request_sock *req)
  */
 
 /* Find the Key structure for an address.  */
-struct tcp_md5sig_key *tcp_md5_do_lookup(struct sock *sk,
+struct tcp_md5sig_key *tcp_md5_do_lookup(const struct sock *sk,
 					 const union tcp_md5_addr *addr,
 					 int family)
 {
@@ -877,7 +877,7 @@ struct tcp_md5sig_key *tcp_md5_do_lookup(struct sock *sk,
 	/* caller either holds rcu_read_lock() or socket lock */
 	md5sig = rcu_dereference_check(tp->md5sig_info,
 				       sock_owned_by_user(sk) ||
-				       lockdep_is_held(&sk->sk_lock.slock));
+				       lockdep_is_held((spinlock_t *)&sk->sk_lock.slock));
 	if (!md5sig)
 		return NULL;
 #if IS_ENABLED(CONFIG_IPV6)
@@ -894,7 +894,7 @@ struct tcp_md5sig_key *tcp_md5_do_lookup(struct sock *sk,
 }
 EXPORT_SYMBOL(tcp_md5_do_lookup);
 
-struct tcp_md5sig_key *tcp_v4_md5_lookup(struct sock *sk,
+struct tcp_md5sig_key *tcp_v4_md5_lookup(const struct sock *sk,
 					 const struct sock *addr_sk)
 {
 	const union tcp_md5_addr *addr;
@@ -1168,7 +1168,8 @@ static bool tcp_v4_inbound_md5_hash(struct sock *sk,
 }
 #endif
 
-static void tcp_v4_init_req(struct request_sock *req, struct sock *sk_listener,
+static void tcp_v4_init_req(struct request_sock *req,
+			    const struct sock *sk_listener,
 			    struct sk_buff *skb)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 85830bb92d04..e0a87c238882 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -362,27 +362,35 @@ void tcp_twsk_destructor(struct sock *sk)
 }
 EXPORT_SYMBOL_GPL(tcp_twsk_destructor);
 
+/* Warning : This function is called without sk_listener being locked.
+ * Be sure to read socket fields once, as their value could change under us.
+ */
 void tcp_openreq_init_rwin(struct request_sock *req,
-			   struct sock *sk, struct dst_entry *dst)
+			   const struct sock *sk_listener,
+			   const struct dst_entry *dst)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
-	struct tcp_sock *tp = tcp_sk(sk);
-	__u8 rcv_wscale;
+	const struct tcp_sock *tp = tcp_sk(sk_listener);
+	u16 user_mss = READ_ONCE(tp->rx_opt.user_mss);
+	int full_space = tcp_full_space(sk_listener);
 	int mss = dst_metric_advmss(dst);
+	u32 window_clamp;
+	__u8 rcv_wscale;
 
-	if (tp->rx_opt.user_mss && tp->rx_opt.user_mss < mss)
-		mss = tp->rx_opt.user_mss;
+	if (user_mss && user_mss < mss)
+		mss = user_mss;
 
+	window_clamp = READ_ONCE(tp->window_clamp);
 	/* Set this up on the first call only */
-	req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW);
+	req->window_clamp = window_clamp ? : dst_metric(dst, RTAX_WINDOW);
 
 	/* limit the window selection if the user enforce a smaller rx buffer */
-	if (sk->sk_userlocks & SOCK_RCVBUF_LOCK &&
-	    (req->window_clamp > tcp_full_space(sk) || req->window_clamp == 0))
-		req->window_clamp = tcp_full_space(sk);
+	if (sk_listener->sk_userlocks & SOCK_RCVBUF_LOCK &&
+	    (req->window_clamp > full_space || req->window_clamp == 0))
+		req->window_clamp = full_space;
 
 	/* tcp_full_space because it is guaranteed to be the first packet */
-	tcp_select_initial_window(tcp_full_space(sk),
+	tcp_select_initial_window(full_space,
 		mss - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0),
 		&req->rcv_wnd,
 		&req->window_clamp,
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 4cd0b50d4e46..53ce6cf55598 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -357,14 +357,10 @@ static void tcp_ecn_clear_syn(struct sock *sk, struct sk_buff *skb)
 }
 
 static void
-tcp_ecn_make_synack(const struct request_sock *req, struct tcphdr *th,
-		    struct sock *sk)
+tcp_ecn_make_synack(const struct request_sock *req, struct tcphdr *th)
 {
-	if (inet_rsk(req)->ecn_ok) {
+	if (inet_rsk(req)->ecn_ok)
 		th->ece = 1;
-		if (tcp_ca_needs_ecn(sk))
-			INET_ECN_xmit(sk);
-	}
 }
 
 /* Set up ECN state for a packet on a ESTABLISHED socket that is about to
@@ -612,12 +608,11 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb,
 }
 
 /* Set up TCP options for SYN-ACKs. */
-static unsigned int tcp_synack_options(struct sock *sk,
-				   struct request_sock *req,
-				   unsigned int mss, struct sk_buff *skb,
-				   struct tcp_out_options *opts,
-				   const struct tcp_md5sig_key *md5,
-				   struct tcp_fastopen_cookie *foc)
+static unsigned int tcp_synack_options(struct request_sock *req,
+				       unsigned int mss, struct sk_buff *skb,
+				       struct tcp_out_options *opts,
+				       const struct tcp_md5sig_key *md5,
+				       struct tcp_fastopen_cookie *foc)
 {
 	struct inet_request_sock *ireq = inet_rsk(req);
 	unsigned int remaining = MAX_TCP_OPTION_SPACE;
@@ -2949,20 +2944,25 @@ int tcp_send_synack(struct sock *sk)
  * Allocate one skb and build a SYNACK packet.
  * @dst is consumed : Caller should not use it again.
  */
-struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
+struct sk_buff *tcp_make_synack(const struct sock *sk, struct dst_entry *dst,
 				struct request_sock *req,
 				struct tcp_fastopen_cookie *foc)
 {
-	struct tcp_out_options opts;
 	struct inet_request_sock *ireq = inet_rsk(req);
-	struct tcp_sock *tp = tcp_sk(sk);
-	struct tcphdr *th;
-	struct sk_buff *skb;
+	const struct tcp_sock *tp = tcp_sk(sk);
 	struct tcp_md5sig_key *md5 = NULL;
+	struct tcp_out_options opts;
+	struct sk_buff *skb;
 	int tcp_header_size;
+	struct tcphdr *th;
+	u16 user_mss;
 	int mss;
 
-	skb = sock_wmalloc(sk, MAX_TCP_HEADER, 1, GFP_ATOMIC);
+	/* sk is a const pointer, because we want to express multiple cpus
+	 * might call us concurrently.
+	 * sock_wmalloc() will change sk->sk_wmem_alloc in an atomic way.
+	 */
+	skb = sock_wmalloc((struct sock *)sk, MAX_TCP_HEADER, 1, GFP_ATOMIC);
 	if (unlikely(!skb)) {
 		dst_release(dst);
 		return NULL;
@@ -2973,8 +2973,9 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 	skb_dst_set(skb, dst);
 
 	mss = dst_metric_advmss(dst);
-	if (tp->rx_opt.user_mss && tp->rx_opt.user_mss < mss)
-		mss = tp->rx_opt.user_mss;
+	user_mss = READ_ONCE(tp->rx_opt.user_mss);
+	if (user_mss && user_mss < mss)
+		mss = user_mss;
 
 	memset(&opts, 0, sizeof(opts));
 #ifdef CONFIG_SYN_COOKIES
@@ -2989,8 +2990,8 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 	md5 = tcp_rsk(req)->af_specific->req_md5_lookup(sk, req_to_sk(req));
 #endif
 	skb_set_hash(skb, tcp_rsk(req)->txhash, PKT_HASH_TYPE_L4);
-	tcp_header_size = tcp_synack_options(sk, req, mss, skb, &opts, md5,
-					     foc) + sizeof(*th);
+	tcp_header_size = tcp_synack_options(req, mss, skb, &opts, md5, foc) +
+			  sizeof(*th);
 
 	skb_push(skb, tcp_header_size);
 	skb_reset_transport_header(skb);
@@ -2999,7 +3000,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 	memset(th, 0, sizeof(struct tcphdr));
 	th->syn = 1;
 	th->ack = 1;
-	tcp_ecn_make_synack(req, th, sk);
+	tcp_ecn_make_synack(req, th);
 	th->source = htons(ireq->ir_num);
 	th->dest = ireq->ir_rmt_port;
 	/* Setting of flags are superfluous here for callers (and ECE is
@@ -3014,7 +3015,7 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
 
 	/* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */
 	th->window = htons(min(req->rcv_wnd, 65535U));
-	tcp_options_write((__be32 *)(th + 1), tp, &opts);
+	tcp_options_write((__be32 *)(th + 1), NULL, &opts);
 	th->doff = (tcp_header_size >> 2);
 	TCP_INC_STATS_BH(sock_net(sk), TCP_MIB_OUTSEGS);
 
@@ -3501,7 +3502,7 @@ void tcp_send_probe0(struct sock *sk)
 				  TCP_RTO_MAX);
 }
 
-int tcp_rtx_synack(struct sock *sk, struct request_sock *req)
+int tcp_rtx_synack(const struct sock *sk, struct request_sock *req)
 {
 	const struct tcp_request_sock_ops *af_ops = tcp_rsk(req)->af_specific;
 	struct flowi fl;