Merge branch 'mptcp-Add-port-parameter-to-ADD_ADDR-option'

Mat Martineau says:

====================
mptcp: Add port parameter to ADD_ADDR option

The ADD_ADDR MPTCP option is used to announce available IP addresses
that a peer may connect to when adding more TCP subflows to an existing
MPTCP connection. There is an optional port number field in that
ADD_ADDR header, and this patch set adds capability for that port number
to be sent and received.

Patches 1, 2, and 4 refactor existing ADD_ADDR code to simplify implementation
of port number support.

Patches 3 and 5 are the main functional changes, for sending and
receiving the port number in the MPTCP ADD_ADDR option.

Patch 6 sends the ADD_ADDR option with port number on a bare TCP ACK,
since the extra length of the option may run in to cases where
sufficient TCP option space is not available on a data packet.

Patch 7 plumbs in port number support for the in-kernel MPTCP path
manager.

Patches 8-11 add some optional debug output and a little more cleanup
refactoring.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/mptcp.h

@@ -46,6 +46,7 @@ struct mptcp_out_options {
 #endif
 	};
 	u8 addr_id;
+	u16 port;
 	u64 ahmac;
 	u8 rm_id;
 	u8 join_id;

net/mptcp/options.c

@@ -242,9 +242,6 @@ static void mptcp_parse_option(const struct sk_buff *skb,
 
 		mp_opt->add_addr = 1;
 		mp_opt->addr_id = *ptr++;
-		pr_debug("ADD_ADDR%s: id=%d, echo=%d",
-			 (mp_opt->family == MPTCP_ADDR_IPVERSION_6) ? "6" : "",
-			 mp_opt->addr_id, mp_opt->echo);
 		if (mp_opt->family == MPTCP_ADDR_IPVERSION_4) {
 			memcpy((u8 *)&mp_opt->addr.s_addr, (u8 *)ptr, 4);
 			ptr += 4;
@@ -269,6 +266,9 @@ static void mptcp_parse_option(const struct sk_buff *skb,
 			mp_opt->ahmac = get_unaligned_be64(ptr);
 			ptr += 8;
 		}
+		pr_debug("ADD_ADDR%s: id=%d, ahmac=%llu, echo=%d, port=%d",
+			 (mp_opt->family == MPTCP_ADDR_IPVERSION_6) ? "6" : "",
+			 mp_opt->addr_id, mp_opt->ahmac, mp_opt->echo, mp_opt->port);
 		break;
 
 	case MPTCPOPT_RM_ADDR:
@@ -587,9 +587,11 @@ static bool mptcp_established_options_add_addr(struct sock *sk, struct sk_buff *
 	unsigned int opt_size = *size;
 	struct mptcp_addr_info saddr;
 	bool echo;
+	bool port;
 	int len;
 
-	if (mptcp_pm_should_add_signal_ipv6(msk) &&
+	if ((mptcp_pm_should_add_signal_ipv6(msk) ||
+	     mptcp_pm_should_add_signal_port(msk)) &&
 	    skb && skb_is_tcp_pure_ack(skb)) {
 		pr_debug("drop other suboptions");
 		opts->suboptions = 0;
@@ -598,10 +600,10 @@ static bool mptcp_established_options_add_addr(struct sock *sk, struct sk_buff *
 	}
 
 	if (!mptcp_pm_should_add_signal(msk) ||
-	    !(mptcp_pm_add_addr_signal(msk, remaining, &saddr, &echo)))
+	    !(mptcp_pm_add_addr_signal(msk, remaining, &saddr, &echo, &port)))
 		return false;
 
-	len = mptcp_add_addr_len(saddr.family, echo);
+	len = mptcp_add_addr_len(saddr.family, echo, port);
 	if (remaining < len)
 		return false;
 
@@ -609,6 +611,8 @@ static bool mptcp_established_options_add_addr(struct sock *sk, struct sk_buff *
 	if (drop_other_suboptions)
 		*size -= opt_size;
 	opts->addr_id = saddr.id;
+	if (port)
+		opts->port = ntohs(saddr.port);
 	if (saddr.family == AF_INET) {
 		opts->suboptions |= OPTION_MPTCP_ADD_ADDR;
 		opts->addr = saddr.addr;
@@ -631,7 +635,8 @@ static bool mptcp_established_options_add_addr(struct sock *sk, struct sk_buff *
 		}
 	}
 #endif
-	pr_debug("addr_id=%d, ahmac=%llu, echo=%d", opts->addr_id, opts->ahmac, echo);
+	pr_debug("addr_id=%d, ahmac=%llu, echo=%d, port=%d",
+		 opts->addr_id, opts->ahmac, echo, opts->port);
 
 	return true;
 }
@@ -1070,44 +1075,66 @@ void mptcp_write_options(__be32 *ptr, const struct tcp_sock *tp,
 	}
 
 mp_capable_done:
-	if (OPTION_MPTCP_ADD_ADDR & opts->suboptions) {
-		if (opts->ahmac)
-			*ptr++ = mptcp_option(MPTCPOPT_ADD_ADDR,
-					      TCPOLEN_MPTCP_ADD_ADDR, 0,
-					      opts->addr_id);
-		else
-			*ptr++ = mptcp_option(MPTCPOPT_ADD_ADDR,
-					      TCPOLEN_MPTCP_ADD_ADDR_BASE,
-					      MPTCP_ADDR_ECHO,
-					      opts->addr_id);
-		memcpy((u8 *)ptr, (u8 *)&opts->addr.s_addr, 4);
-		ptr += 1;
+	if ((OPTION_MPTCP_ADD_ADDR
+#if IS_ENABLED(CONFIG_MPTCP_IPV6)
+	     | OPTION_MPTCP_ADD_ADDR6
+#endif
+	    ) & opts->suboptions) {
+		u8 len = TCPOLEN_MPTCP_ADD_ADDR_BASE;
+		u8 echo = MPTCP_ADDR_ECHO;
+
+#if IS_ENABLED(CONFIG_MPTCP_IPV6)
+		if (OPTION_MPTCP_ADD_ADDR6 & opts->suboptions)
+			len = TCPOLEN_MPTCP_ADD_ADDR6_BASE;
+#endif
+
+		if (opts->port)
+			len += TCPOLEN_MPTCP_PORT_LEN;
+
 		if (opts->ahmac) {
-			put_unaligned_be64(opts->ahmac, ptr);
-			ptr += 2;
+			len += sizeof(opts->ahmac);
+			echo = 0;
 		}
-	}
 
+		*ptr++ = mptcp_option(MPTCPOPT_ADD_ADDR,
+				      len, echo, opts->addr_id);
+		if (OPTION_MPTCP_ADD_ADDR & opts->suboptions) {
+			memcpy((u8 *)ptr, (u8 *)&opts->addr.s_addr, 4);
+			ptr += 1;
+		}
 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
-	if (OPTION_MPTCP_ADD_ADDR6 & opts->suboptions) {
-		if (opts->ahmac)
-			*ptr++ = mptcp_option(MPTCPOPT_ADD_ADDR,
-					      TCPOLEN_MPTCP_ADD_ADDR6, 0,
-					      opts->addr_id);
-		else
-			*ptr++ = mptcp_option(MPTCPOPT_ADD_ADDR,
-					      TCPOLEN_MPTCP_ADD_ADDR6_BASE,
-					      MPTCP_ADDR_ECHO,
-					      opts->addr_id);
-		memcpy((u8 *)ptr, opts->addr6.s6_addr, 16);
-		ptr += 4;
-		if (opts->ahmac) {
-			put_unaligned_be64(opts->ahmac, ptr);
-			ptr += 2;
+		else if (OPTION_MPTCP_ADD_ADDR6 & opts->suboptions) {
+			memcpy((u8 *)ptr, opts->addr6.s6_addr, 16);
+			ptr += 4;
 		}
-	}
 #endif
 
+		if (!opts->port) {
+			if (opts->ahmac) {
+				put_unaligned_be64(opts->ahmac, ptr);
+				ptr += 2;
+			}
+		} else {
+			if (opts->ahmac) {
+				u8 *bptr = (u8 *)ptr;
+
+				put_unaligned_be16(opts->port, bptr);
+				bptr += 2;
+				put_unaligned_be64(opts->ahmac, bptr);
+				bptr += 8;
+				put_unaligned_be16(TCPOPT_NOP << 8 |
+						   TCPOPT_NOP, bptr);
+
+				ptr += 3;
+			} else {
+				put_unaligned_be32(opts->port << 16 |
+						   TCPOPT_NOP << 8 |
+						   TCPOPT_NOP, ptr);
+				ptr += 1;
+			}
+		}
+	}
+
 	if (OPTION_MPTCP_RM_ADDR & opts->suboptions) {
 		*ptr++ = mptcp_option(MPTCPOPT_RM_ADDR,
 				      TCPOLEN_MPTCP_RM_ADDR_BASE,

net/mptcp/pm.c

@@ -14,28 +14,43 @@
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
 			   const struct mptcp_addr_info *addr,
-			   bool echo)
+			   bool echo, bool port)
 {
-	u8 add_addr = READ_ONCE(msk->pm.add_addr_signal);
+	u8 add_addr = READ_ONCE(msk->pm.addr_signal);
 
 	pr_debug("msk=%p, local_id=%d", msk, addr->id);
 
+	if (add_addr) {
+		pr_warn("addr_signal error, add_addr=%d", add_addr);
+		return -EINVAL;
+	}
+
 	msk->pm.local = *addr;
 	add_addr |= BIT(MPTCP_ADD_ADDR_SIGNAL);
 	if (echo)
 		add_addr |= BIT(MPTCP_ADD_ADDR_ECHO);
 	if (addr->family == AF_INET6)
 		add_addr |= BIT(MPTCP_ADD_ADDR_IPV6);
-	WRITE_ONCE(msk->pm.add_addr_signal, add_addr);
+	if (port)
+		add_addr |= BIT(MPTCP_ADD_ADDR_PORT);
+	WRITE_ONCE(msk->pm.addr_signal, add_addr);
 	return 0;
 }
 
 int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id)
 {
+	u8 rm_addr = READ_ONCE(msk->pm.addr_signal);
+
 	pr_debug("msk=%p, local_id=%d", msk, local_id);
 
+	if (rm_addr) {
+		pr_warn("addr_signal error, rm_addr=%d", rm_addr);
+		return -EINVAL;
+	}
+
 	msk->pm.rm_id = local_id;
-	WRITE_ONCE(msk->pm.rm_addr_signal, true);
+	rm_addr |= BIT(MPTCP_RM_ADDR_SIGNAL);
+	WRITE_ONCE(msk->pm.addr_signal, rm_addr);
 	return 0;
 }
 
@@ -156,7 +171,7 @@ void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
 	spin_lock_bh(&pm->lock);
 
 	if (!READ_ONCE(pm->accept_addr)) {
-		mptcp_pm_announce_addr(msk, addr, true);
+		mptcp_pm_announce_addr(msk, addr, true, addr->port);
 		mptcp_pm_add_addr_send_ack(msk);
 	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
 		pm->remote = *addr;
@@ -167,7 +182,8 @@ void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
 
 void mptcp_pm_add_addr_send_ack(struct mptcp_sock *msk)
 {
-	if (!mptcp_pm_should_add_signal_ipv6(msk))
+	if (!mptcp_pm_should_add_signal_ipv6(msk) &&
+	    !mptcp_pm_should_add_signal_port(msk))
 		return;
 
 	mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_SEND_ACK);
@@ -188,7 +204,7 @@ void mptcp_pm_rm_addr_received(struct mptcp_sock *msk, u8 rm_id)
 /* path manager helpers */
 
 bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
-			      struct mptcp_addr_info *saddr, bool *echo)
+			      struct mptcp_addr_info *saddr, bool *echo, bool *port)
 {
 	int ret = false;
 
@@ -199,12 +215,13 @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 		goto out_unlock;
 
 	*echo = mptcp_pm_should_add_signal_echo(msk);
+	*port = mptcp_pm_should_add_signal_port(msk);
 
-	if (remaining < mptcp_add_addr_len(msk->pm.local.family, *echo))
+	if (remaining < mptcp_add_addr_len(msk->pm.local.family, *echo, *port))
 		goto out_unlock;
 
 	*saddr = msk->pm.local;
-	WRITE_ONCE(msk->pm.add_addr_signal, 0);
+	WRITE_ONCE(msk->pm.addr_signal, 0);
 	ret = true;
 
 out_unlock:
@@ -227,7 +244,7 @@ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 		goto out_unlock;
 
 	*rm_id = msk->pm.rm_id;
-	WRITE_ONCE(msk->pm.rm_addr_signal, false);
+	WRITE_ONCE(msk->pm.addr_signal, 0);
 	ret = true;
 
 out_unlock:
@@ -248,8 +265,7 @@ void mptcp_pm_data_init(struct mptcp_sock *msk)
 	msk->pm.subflows = 0;
 	msk->pm.rm_id = 0;
 	WRITE_ONCE(msk->pm.work_pending, false);
-	WRITE_ONCE(msk->pm.add_addr_signal, 0);
-	WRITE_ONCE(msk->pm.rm_addr_signal, false);
+	WRITE_ONCE(msk->pm.addr_signal, 0);
 	WRITE_ONCE(msk->pm.accept_addr, false);
 	WRITE_ONCE(msk->pm.accept_subflow, false);
 	msk->pm.status = 0;

net/mptcp/pm_netlink.c

@@ -227,7 +227,7 @@ static void mptcp_pm_add_timer(struct timer_list *timer)
 
 	if (!mptcp_pm_should_add_signal(msk)) {
 		pr_debug("retransmit ADD_ADDR id=%d", entry->addr.id);
-		mptcp_pm_announce_addr(msk, &entry->addr, false);
+		mptcp_pm_announce_addr(msk, &entry->addr, false, entry->addr.port);
 		mptcp_pm_add_addr_send_ack(msk);
 		entry->retrans_times++;
 	}
@@ -313,7 +313,7 @@ static void mptcp_pm_create_subflow_or_signal_addr(struct mptcp_sock *msk)
 	struct mptcp_pm_addr_entry *local;
 	struct pm_nl_pernet *pernet;
 
-	pernet = net_generic(sock_net((struct sock *)msk), pm_nl_pernet_id);
+	pernet = net_generic(sock_net(sk), pm_nl_pernet_id);
 
 	pr_debug("local %d:%d signal %d:%d subflows %d:%d\n",
 		 msk->pm.local_addr_used, msk->pm.local_addr_max,
@@ -328,7 +328,7 @@ static void mptcp_pm_create_subflow_or_signal_addr(struct mptcp_sock *msk)
 		if (local) {
 			if (mptcp_pm_alloc_anno_list(msk, local)) {
 				msk->pm.add_addr_signaled++;
-				mptcp_pm_announce_addr(msk, &local->addr, false);
+				mptcp_pm_announce_addr(msk, &local->addr, false, local->addr.port);
 				mptcp_pm_nl_add_addr_send_ack(msk);
 			}
 		} else {
@@ -376,6 +376,7 @@ void mptcp_pm_nl_add_addr_received(struct mptcp_sock *msk)
 	struct sock *sk = (struct sock *)msk;
 	struct mptcp_addr_info remote;
 	struct mptcp_addr_info local;
+	bool use_port = false;
 
 	pr_debug("accepted %d:%d remote family %d",
 		 msk->pm.add_addr_accepted, msk->pm.add_addr_accept_max,
@@ -392,14 +393,16 @@ void mptcp_pm_nl_add_addr_received(struct mptcp_sock *msk)
 	remote = msk->pm.remote;
 	if (!remote.port)
 		remote.port = sk->sk_dport;
+	else
+		use_port = true;
 	memset(&local, 0, sizeof(local));
 	local.family = remote.family;
 
 	spin_unlock_bh(&msk->pm.lock);
-	__mptcp_subflow_connect((struct sock *)msk, &local, &remote);
+	__mptcp_subflow_connect(sk, &local, &remote);
 	spin_lock_bh(&msk->pm.lock);
 
-	mptcp_pm_announce_addr(msk, &remote, true);
+	mptcp_pm_announce_addr(msk, &remote, true, use_port);
 	mptcp_pm_nl_add_addr_send_ack(msk);
 }
 
@@ -407,7 +410,8 @@ void mptcp_pm_nl_add_addr_send_ack(struct mptcp_sock *msk)
 {
 	struct mptcp_subflow_context *subflow;
 
-	if (!mptcp_pm_should_add_signal_ipv6(msk))
+	if (!mptcp_pm_should_add_signal_ipv6(msk) &&
+	    !mptcp_pm_should_add_signal_port(msk))
 		return;
 
 	__mptcp_flush_join_list(msk);
@@ -417,15 +421,22 @@ void mptcp_pm_nl_add_addr_send_ack(struct mptcp_sock *msk)
 		u8 add_addr;
 
 		spin_unlock_bh(&msk->pm.lock);
-		pr_debug("send ack for add_addr6");
+		if (mptcp_pm_should_add_signal_ipv6(msk))
+			pr_debug("send ack for add_addr6");
+		if (mptcp_pm_should_add_signal_port(msk))
+			pr_debug("send ack for add_addr_port");
+
 		lock_sock(ssk);
 		tcp_send_ack(ssk);
 		release_sock(ssk);
 		spin_lock_bh(&msk->pm.lock);
 
-		add_addr = READ_ONCE(msk->pm.add_addr_signal);
-		add_addr &= ~BIT(MPTCP_ADD_ADDR_IPV6);
-		WRITE_ONCE(msk->pm.add_addr_signal, add_addr);
+		add_addr = READ_ONCE(msk->pm.addr_signal);
+		if (mptcp_pm_should_add_signal_ipv6(msk))
+			add_addr &= ~BIT(MPTCP_ADD_ADDR_IPV6);
+		if (mptcp_pm_should_add_signal_port(msk))
+			add_addr &= ~BIT(MPTCP_ADD_ADDR_PORT);
+		WRITE_ONCE(msk->pm.addr_signal, add_addr);
 	}
 }
 

net/mptcp/protocol.h

@@ -49,14 +49,14 @@
 #define TCPOLEN_MPTCP_DSS_MAP64		14
 #define TCPOLEN_MPTCP_DSS_CHECKSUM	2
 #define TCPOLEN_MPTCP_ADD_ADDR		16
-#define TCPOLEN_MPTCP_ADD_ADDR_PORT	18
+#define TCPOLEN_MPTCP_ADD_ADDR_PORT	20
 #define TCPOLEN_MPTCP_ADD_ADDR_BASE	8
-#define TCPOLEN_MPTCP_ADD_ADDR_BASE_PORT	10
+#define TCPOLEN_MPTCP_ADD_ADDR_BASE_PORT	12
 #define TCPOLEN_MPTCP_ADD_ADDR6		28
-#define TCPOLEN_MPTCP_ADD_ADDR6_PORT	30
+#define TCPOLEN_MPTCP_ADD_ADDR6_PORT	32
 #define TCPOLEN_MPTCP_ADD_ADDR6_BASE	20
-#define TCPOLEN_MPTCP_ADD_ADDR6_BASE_PORT	22
-#define TCPOLEN_MPTCP_PORT_LEN		2
+#define TCPOLEN_MPTCP_ADD_ADDR6_BASE_PORT	24
+#define TCPOLEN_MPTCP_PORT_LEN		4
 #define TCPOLEN_MPTCP_RM_ADDR_BASE	4
 
 /* MPTCP MP_JOIN flags */
@@ -168,10 +168,12 @@ enum mptcp_pm_status {
 	MPTCP_PM_SUBFLOW_ESTABLISHED,
 };
 
-enum mptcp_add_addr_status {
+enum mptcp_addr_signal_status {
 	MPTCP_ADD_ADDR_SIGNAL,
 	MPTCP_ADD_ADDR_ECHO,
 	MPTCP_ADD_ADDR_IPV6,
+	MPTCP_ADD_ADDR_PORT,
+	MPTCP_RM_ADDR_SIGNAL,
 };
 
 struct mptcp_pm_data {
@@ -181,8 +183,7 @@ struct mptcp_pm_data {
 
 	spinlock_t	lock;		/*protects the whole PM data */
 
-	u8		add_addr_signal;
-	bool		rm_addr_signal;
+	u8		addr_signal;
 	bool		server_side;
 	bool		work_pending;
 	bool		accept_addr;
@@ -551,40 +552,51 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk,
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
 			   const struct mptcp_addr_info *addr,
-			   bool echo);
+			   bool echo, bool port);
 int mptcp_pm_remove_addr(struct mptcp_sock *msk, u8 local_id);
 int mptcp_pm_remove_subflow(struct mptcp_sock *msk, u8 local_id);
 
 static inline bool mptcp_pm_should_add_signal(struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.add_addr_signal) & BIT(MPTCP_ADD_ADDR_SIGNAL);
+	return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_ADD_ADDR_SIGNAL);
 }
 
 static inline bool mptcp_pm_should_add_signal_echo(struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.add_addr_signal) & BIT(MPTCP_ADD_ADDR_ECHO);
+	return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_ADD_ADDR_ECHO);
 }
 
 static inline bool mptcp_pm_should_add_signal_ipv6(struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.add_addr_signal) & BIT(MPTCP_ADD_ADDR_IPV6);
+	return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_ADD_ADDR_IPV6);
+}
+
+static inline bool mptcp_pm_should_add_signal_port(struct mptcp_sock *msk)
+{
+	return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_ADD_ADDR_PORT);
 }
 
 static inline bool mptcp_pm_should_rm_signal(struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.rm_addr_signal);
+	return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_RM_ADDR_SIGNAL);
 }
 
-static inline unsigned int mptcp_add_addr_len(int family, bool echo)
+static inline unsigned int mptcp_add_addr_len(int family, bool echo, bool port)
 {
-	if (family == AF_INET)
-		return echo ? TCPOLEN_MPTCP_ADD_ADDR_BASE
-			    : TCPOLEN_MPTCP_ADD_ADDR;
-	return echo ? TCPOLEN_MPTCP_ADD_ADDR6_BASE : TCPOLEN_MPTCP_ADD_ADDR6;
+	u8 len = TCPOLEN_MPTCP_ADD_ADDR_BASE;
+
+	if (family == AF_INET6)
+		len = TCPOLEN_MPTCP_ADD_ADDR6_BASE;
+	if (!echo)
+		len += MPTCPOPT_THMAC_LEN;
+	if (port)
+		len += TCPOLEN_MPTCP_PORT_LEN;
+
+	return len;
 }
 
 bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
-			      struct mptcp_addr_info *saddr, bool *echo);
+			      struct mptcp_addr_info *saddr, bool *echo, bool *port);
 bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 			     u8 *rm_id);
 int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);