Merge branch 'tipc-separate-link-and-aggregation'

Jon Maloy says:

====================
tipc: separate link and link aggregation layer

This is the first batch of a longer series that has two main objectives:

o Finer lock granularity during message sending and reception,
  especially regarding usage of the node spinlock.

o Better separation between the link layer implementation and the link
  aggregation layer, represented by node.c::struct tipc_node.

Hopefully these changes also make this part of code somewhat easier
to comprehend and maintain.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

net/tipc/bcast.c

@@ -316,6 +316,29 @@ void tipc_bclink_update_link_state(struct tipc_node *n_ptr,
 	}
 }
 
+void tipc_bclink_sync_state(struct tipc_node *n, struct tipc_msg *hdr)
+{
+	u16 last = msg_last_bcast(hdr);
+	int mtyp = msg_type(hdr);
+
+	if (unlikely(msg_user(hdr) != LINK_PROTOCOL))
+		return;
+	if (mtyp == STATE_MSG) {
+		tipc_bclink_update_link_state(n, last);
+		return;
+	}
+	/* Compatibility: older nodes don't know BCAST_PROTOCOL synchronization,
+	 * and transfer synch info in LINK_PROTOCOL messages.
+	 */
+	if (tipc_node_is_up(n))
+		return;
+	if ((mtyp != RESET_MSG) && (mtyp != ACTIVATE_MSG))
+		return;
+	n->bclink.last_sent = last;
+	n->bclink.last_in = last;
+	n->bclink.oos_state = 0;
+}
+
 /**
  * bclink_peek_nack - monitor retransmission requests sent by other nodes
  *
@@ -358,10 +381,9 @@ int tipc_bclink_xmit(struct net *net, struct sk_buff_head *list)
 
 	/* Prepare clone of message for local node */
 	skb = tipc_msg_reassemble(list);
-	if (unlikely(!skb)) {
-		__skb_queue_purge(list);
+	if (unlikely(!skb))
 		return -EHOSTUNREACH;
-	}
+
 	/* Broadcast to all nodes */
 	if (likely(bclink)) {
 		tipc_bclink_lock(net);
@@ -413,7 +435,7 @@ static void bclink_accept_pkt(struct tipc_node *node, u32 seqno)
 	 * all nodes in the cluster don't ACK at the same time
 	 */
 	if (((seqno - tn->own_addr) % TIPC_MIN_LINK_WIN) == 0) {
-		tipc_link_proto_xmit(node->active_links[node->addr & 1],
+		tipc_link_proto_xmit(node_active_link(node, node->addr),
 				     STATE_MSG, 0, 0, 0, 0);
 		tn->bcl->stats.sent_acks++;
 	}
@@ -925,7 +947,6 @@ int tipc_bclink_init(struct net *net)
 	tipc_link_set_queue_limits(bcl, BCLINK_WIN_DEFAULT);
 	bcl->bearer_id = MAX_BEARERS;
 	rcu_assign_pointer(tn->bearer_list[MAX_BEARERS], &bcbearer->bearer);
-	bcl->state = WORKING_WORKING;
 	bcl->pmsg = (struct tipc_msg *)&bcl->proto_msg;
 	msg_set_prevnode(bcl->pmsg, tn->own_addr);
 	strlcpy(bcl->name, tipc_bclink_name, TIPC_MAX_LINK_NAME);

net/tipc/bcast.h

@@ -133,5 +133,6 @@ void tipc_bclink_wakeup_users(struct net *net);
 int tipc_nl_add_bc_link(struct net *net, struct tipc_nl_msg *msg);
 int tipc_nl_bc_link_set(struct net *net, struct nlattr *attrs[]);
 void tipc_bclink_input(struct net *net);
+void tipc_bclink_sync_state(struct tipc_node *n, struct tipc_msg *msg);
 
 #endif

net/tipc/bearer.c

@@ -470,6 +470,32 @@ void tipc_bearer_send(struct net *net, u32 bearer_id, struct sk_buff *buf,
 	rcu_read_unlock();
 }
 
+/* tipc_bearer_xmit() -send buffer to destination over bearer
+ */
+void tipc_bearer_xmit(struct net *net, u32 bearer_id,
+		      struct sk_buff_head *xmitq,
+		      struct tipc_media_addr *dst)
+{
+	struct tipc_net *tn = net_generic(net, tipc_net_id);
+	struct tipc_bearer *b;
+	struct sk_buff *skb, *tmp;
+
+	if (skb_queue_empty(xmitq))
+		return;
+
+	rcu_read_lock();
+	b = rcu_dereference_rtnl(tn->bearer_list[bearer_id]);
+	if (likely(b)) {
+		skb_queue_walk_safe(xmitq, skb, tmp) {
+			__skb_dequeue(xmitq);
+			b->media->send_msg(net, skb, b, dst);
+			/* Until we remove cloning in tipc_l2_send_msg(): */
+			kfree_skb(skb);
+		}
+	}
+	rcu_read_unlock();
+}
+
 /**
  * tipc_l2_rcv_msg - handle incoming TIPC message from an interface
  * @buf: the received packet

net/tipc/bearer.h

@@ -217,5 +217,8 @@ void tipc_bearer_cleanup(void);
 void tipc_bearer_stop(struct net *net);
 void tipc_bearer_send(struct net *net, u32 bearer_id, struct sk_buff *buf,
 		      struct tipc_media_addr *dest);
+void tipc_bearer_xmit(struct net *net, u32 bearer_id,
+		      struct sk_buff_head *xmitq,
+		      struct tipc_media_addr *dst);
 
 #endif	/* _TIPC_BEARER_H */

net/tipc/core.h

@@ -129,6 +129,11 @@ static inline int less(u16 left, u16 right)
 	return less_eq(left, right) && (mod(right) != mod(left));
 }
 
+static inline int in_range(u16 val, u16 min, u16 max)
+{
+	return !less(val, min) && !more(val, max);
+}
+
 #ifdef CONFIG_SYSCTL
 int tipc_register_sysctl(void);
 void tipc_unregister_sysctl(void);

net/tipc/discover.c

@@ -35,7 +35,7 @@
  */
 
 #include "core.h"
-#include "link.h"
+#include "node.h"
 #include "discover.h"
 
 /* min delay during bearer start up */
@@ -125,7 +125,6 @@ void tipc_disc_rcv(struct net *net, struct sk_buff *buf,
 {
 	struct tipc_net *tn = net_generic(net, tipc_net_id);
 	struct tipc_node *node;
-	struct tipc_link *link;
 	struct tipc_media_addr maddr;
 	struct sk_buff *rbuf;
 	struct tipc_msg *msg = buf_msg(buf);
@@ -170,13 +169,10 @@ void tipc_disc_rcv(struct net *net, struct sk_buff *buf,
 		return;
 	tipc_node_lock(node);
 	node->capabilities = caps;
-	link = node->links[bearer->identity];
 
 	/* Prepare to validate requesting node's signature and media address */
 	sign_match = (signature == node->signature);
-	addr_match = link && !memcmp(&link->media_addr, &maddr, sizeof(maddr));
-	link_up = link && tipc_link_is_up(link);
-
+	tipc_node_check_dest(node, bearer, &link_up, &addr_match, &maddr);
 
 	/* These three flags give us eight permutations: */
 
@@ -239,16 +235,8 @@ void tipc_disc_rcv(struct net *net, struct sk_buff *buf,
 	if (accept_sign)
 		node->signature = signature;
 
-	if (accept_addr) {
-		if (!link)
-			link = tipc_link_create(node, bearer, &maddr);
-		if (link) {
-			memcpy(&link->media_addr, &maddr, sizeof(maddr));
-			tipc_link_reset(link);
-		} else {
+	if (accept_addr && !tipc_node_update_dest(node, bearer, &maddr))
 		respond = false;
-		}
-	}
 
 	/* Send response, if necessary */
 	if (respond && (mtyp == DSC_REQ_MSG)) {

net/tipc/link.h

@@ -49,19 +49,21 @@
  */
 #define INVALID_LINK_SEQ 0x10000
 
-/* Link working states
+
+/* Link endpoint receive states
  */
-#define WORKING_WORKING 560810u
-#define WORKING_UNKNOWN 560811u
-#define RESET_UNKNOWN   560812u
-#define RESET_RESET     560813u
+enum {
+	TIPC_LINK_OPEN,
+	TIPC_LINK_BLOCKED,
+	TIPC_LINK_TUNNEL
+};
 
-/* Link endpoint execution states
+/* Events returned from link at packet reception or at timeout
  */
-#define LINK_STARTED     0x0001
-#define LINK_STOPPED     0x0002
-#define LINK_SYNCHING    0x0004
-#define LINK_FAILINGOVER 0x0008
+enum {
+	TIPC_LINK_UP_EVT       = 1,
+	TIPC_LINK_DOWN_EVT     = (1 << 1)
+};
 
 /* Starting value for maximum packet size negotiation on unicast links
  * (unless bearer MTU is less)
@@ -106,7 +108,6 @@ struct tipc_stats {
  * @timer: link timer
  * @owner: pointer to peer node
  * @refcnt: reference counter for permanent references (owner node & timer)
- * @flags: execution state flags for link endpoint instance
  * @peer_session: link session # being used by peer end of link
  * @peer_bearer_id: bearer id used by link's peer endpoint
  * @bearer_id: local bearer id used by link
@@ -119,6 +120,7 @@ struct tipc_stats {
  * @pmsg: convenience pointer to "proto_msg" field
  * @priority: current link priority
  * @net_plane: current link network plane ('A' through 'H')
+ * @exec_mode: transmit/receive mode for link endpoint instance
  * @backlog_limit: backlog queue congestion thresholds (indexed by importance)
  * @exp_msg_count: # of tunnelled messages expected during link changeover
  * @reset_rcv_checkpt: seq # of last acknowledged message at time of link reset
@@ -144,12 +146,9 @@ struct tipc_link {
 	u32 addr;
 	char name[TIPC_MAX_LINK_NAME];
 	struct tipc_media_addr media_addr;
-	struct timer_list timer;
 	struct tipc_node *owner;
-	struct kref ref;
 
 	/* Management and link supervision data */
-	unsigned int flags;
 	u32 peer_session;
 	u32 peer_bearer_id;
 	u32 bearer_id;
@@ -165,6 +164,7 @@ struct tipc_link {
 	struct tipc_msg *pmsg;
 	u32 priority;
 	char net_plane;
+	u8 exec_mode;
 	u16 synch_point;
 
 	/* Failover */
@@ -192,8 +192,8 @@ struct tipc_link {
 	u16 rcv_nxt;
 	u32 rcv_unacked;
 	struct sk_buff_head deferdq;
-	struct sk_buff_head inputq;
-	struct sk_buff_head namedq;
+	struct sk_buff_head *inputq;
+	struct sk_buff_head *namedq;
 
 	/* Congestion handling */
 	struct sk_buff_head wakeupq;
@@ -207,9 +207,11 @@ struct tipc_link {
 
 struct tipc_port;
 
-struct tipc_link *tipc_link_create(struct tipc_node *n_ptr,
-			      struct tipc_bearer *b_ptr,
-			      const struct tipc_media_addr *media_addr);
+struct tipc_link *tipc_link_create(struct tipc_node *n,
+				   struct tipc_bearer *b,
+				   const struct tipc_media_addr *maddr,
+				   struct sk_buff_head *inputq,
+				   struct sk_buff_head *namedq);
 void tipc_link_delete(struct tipc_link *link);
 void tipc_link_delete_list(struct net *net, unsigned int bearer_id);
 void tipc_link_failover_send_queue(struct tipc_link *l_ptr);
@@ -221,12 +223,11 @@ void tipc_link_purge_queues(struct tipc_link *l_ptr);
 void tipc_link_purge_backlog(struct tipc_link *l);
 void tipc_link_reset_all(struct tipc_node *node);
 void tipc_link_reset(struct tipc_link *l_ptr);
-int tipc_link_xmit_skb(struct net *net, struct sk_buff *skb, u32 dest,
-		       u32 selector);
-int tipc_link_xmit(struct net *net, struct sk_buff_head *list, u32 dest,
-		   u32 selector);
+void tipc_link_activate(struct tipc_link *link);
 int __tipc_link_xmit(struct net *net, struct tipc_link *link,
 		     struct sk_buff_head *list);
+int tipc_link_xmit(struct tipc_link *link,	struct sk_buff_head *list,
+		   struct sk_buff_head *xmitq);
 void tipc_link_proto_xmit(struct tipc_link *l_ptr, u32 msg_typ, int prob,
 			  u32 gap, u32 tolerance, u32 priority);
 void tipc_link_push_packets(struct tipc_link *l_ptr);
@@ -243,33 +244,12 @@ int tipc_nl_link_set(struct sk_buff *skb, struct genl_info *info);
 int tipc_nl_link_reset_stats(struct sk_buff *skb, struct genl_info *info);
 int tipc_nl_parse_link_prop(struct nlattr *prop, struct nlattr *props[]);
 void link_prepare_wakeup(struct tipc_link *l);
-
+int tipc_link_timeout(struct tipc_link *l, struct sk_buff_head *xmitq);
+int tipc_link_rcv(struct tipc_link *l, struct sk_buff *skb,
+		  struct sk_buff_head *xmitq);
 static inline u32 link_own_addr(struct tipc_link *l)
 {
 	return msg_prevnode(l->pmsg);
 }
 
-/*
- * Link status checking routines
- */
-static inline int link_working_working(struct tipc_link *l_ptr)
-{
-	return l_ptr->state == WORKING_WORKING;
-}
-
-static inline int link_working_unknown(struct tipc_link *l_ptr)
-{
-	return l_ptr->state == WORKING_UNKNOWN;
-}
-
-static inline int link_reset_unknown(struct tipc_link *l_ptr)
-{
-	return l_ptr->state == RESET_UNKNOWN;
-}
-
-static inline int link_reset_reset(struct tipc_link *l_ptr)
-{
-	return l_ptr->state == RESET_RESET;
-}
-
 #endif

net/tipc/msg.h

@@ -38,6 +38,7 @@
 #define _TIPC_MSG_H
 
 #include <linux/tipc.h>
+#include "core.h"
 
 /*
  * Constants and routines used to read and write TIPC payload message headers
@@ -658,12 +659,12 @@ static inline void msg_set_link_selector(struct tipc_msg *m, u32 n)
 /*
  * Word 5
  */
-static inline u32 msg_session(struct tipc_msg *m)
+static inline u16 msg_session(struct tipc_msg *m)
 {
 	return msg_bits(m, 5, 16, 0xffff);
 }
 
-static inline void msg_set_session(struct tipc_msg *m, u32 n)
+static inline void msg_set_session(struct tipc_msg *m, u16 n)
 {
 	msg_set_bits(m, 5, 16, 0xffff, n);
 }
@@ -766,6 +767,22 @@ static inline void msg_set_link_tolerance(struct tipc_msg *m, u32 n)
 	msg_set_bits(m, 9, 0, 0xffff, n);
 }
 
+static inline bool msg_is_traffic(struct tipc_msg *m)
+{
+	if (likely(msg_user(m) != LINK_PROTOCOL))
+		return true;
+	if ((msg_type(m) == RESET_MSG) || (msg_type(m) == ACTIVATE_MSG))
+		return false;
+	return true;
+}
+
+static inline bool msg_peer_is_up(struct tipc_msg *m)
+{
+	if (likely(msg_is_traffic(m)))
+		return false;
+	return msg_redundant_link(m);
+}
+
 struct sk_buff *tipc_buf_acquire(u32 size);
 bool tipc_msg_validate(struct sk_buff *skb);
 bool tipc_msg_reverse(u32 own_addr, struct sk_buff *buf, u32 *dnode,
@@ -879,4 +896,36 @@ static inline bool tipc_skb_queue_tail(struct sk_buff_head *list,
 	return rv;
 }
 
+/* tipc_skb_queue_sorted(); sort pkt into list according to sequence number
+ * @list: list to be appended to
+ * @skb: buffer to add
+ * Returns true if queue should treated further, otherwise false
+ */
+static inline bool __tipc_skb_queue_sorted(struct sk_buff_head *list,
+					   struct sk_buff *skb)
+{
+	struct sk_buff *_skb, *tmp;
+	struct tipc_msg *hdr = buf_msg(skb);
+	u16 seqno = msg_seqno(hdr);
+
+	if (skb_queue_empty(list) || (msg_user(hdr) == LINK_PROTOCOL)) {
+		__skb_queue_head(list, skb);
+		return true;
+	}
+	if (likely(less(seqno, buf_seqno(skb_peek(list))))) {
+		__skb_queue_head(list, skb);
+		return true;
+	}
+	if (!more(seqno, buf_seqno(skb_peek_tail(list)))) {
+		skb_queue_walk_safe(list, _skb, tmp) {
+			if (likely(less(seqno, buf_seqno(_skb)))) {
+				__skb_queue_before(list, _skb, skb);
+				return true;
+			}
+		}
+	}
+	__skb_queue_tail(list, skb);
+	return false;
+}
+
 #endif

net/tipc/name_distr.c

@@ -96,13 +96,13 @@ void named_cluster_distribute(struct net *net, struct sk_buff *skb)
 		dnode = node->addr;
 		if (in_own_node(net, dnode))
 			continue;
-		if (!tipc_node_active_links(node))
+		if (!tipc_node_is_up(node))
 			continue;
 		oskb = pskb_copy(skb, GFP_ATOMIC);
 		if (!oskb)
 			break;
 		msg_set_destnode(buf_msg(oskb), dnode);
-		tipc_link_xmit_skb(net, oskb, dnode, dnode);
+		tipc_node_xmit_skb(net, oskb, dnode, dnode);
 	}
 	rcu_read_unlock();
 
@@ -223,7 +223,7 @@ void tipc_named_node_up(struct net *net, u32 dnode)
 			 &tn->nametbl->publ_list[TIPC_ZONE_SCOPE]);
 	rcu_read_unlock();
 
-	tipc_link_xmit(net, &head, dnode, dnode);
+	tipc_node_xmit(net, &head, dnode, dnode);
 }
 
 static void tipc_publ_subscribe(struct net *net, struct publication *publ,

net/tipc/node.c

@@ -40,10 +40,13 @@
 #include "name_distr.h"
 #include "socket.h"
 #include "bcast.h"
+#include "discover.h"
 
 static void node_lost_contact(struct tipc_node *n_ptr);
 static void node_established_contact(struct tipc_node *n_ptr);
 static void tipc_node_delete(struct tipc_node *node);
+static void tipc_node_timeout(unsigned long data);
+static void tipc_node_fsm_evt(struct tipc_node *n, int evt);
 
 struct tipc_sock_conn {
 	u32 port;
@@ -132,6 +135,7 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr)
 	INIT_LIST_HEAD(&n_ptr->list);
 	INIT_LIST_HEAD(&n_ptr->publ_list);
 	INIT_LIST_HEAD(&n_ptr->conn_sks);
+	skb_queue_head_init(&n_ptr->bclink.namedq);
 	__skb_queue_head_init(&n_ptr->bclink.deferdq);
 	hlist_add_head_rcu(&n_ptr->hash, &tn->node_htable[tipc_hashfn(addr)]);
 	list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
@@ -139,14 +143,32 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr)
 			break;
 	}
 	list_add_tail_rcu(&n_ptr->list, &temp_node->list);
-	n_ptr->action_flags = TIPC_WAIT_PEER_LINKS_DOWN;
+	n_ptr->state = SELF_DOWN_PEER_LEAVING;
 	n_ptr->signature = INVALID_NODE_SIG;
+	n_ptr->active_links[0] = INVALID_BEARER_ID;
+	n_ptr->active_links[1] = INVALID_BEARER_ID;
 	tipc_node_get(n_ptr);
+	setup_timer(&n_ptr->timer, tipc_node_timeout, (unsigned long)n_ptr);
+	n_ptr->keepalive_intv = U32_MAX;
 exit:
 	spin_unlock_bh(&tn->node_list_lock);
 	return n_ptr;
 }
 
+static void tipc_node_calculate_timer(struct tipc_node *n, struct tipc_link *l)
+{
+	unsigned long tol = l->tolerance;
+	unsigned long intv = ((tol / 4) > 500) ? 500 : tol / 4;
+	unsigned long keepalive_intv = msecs_to_jiffies(intv);
+
+	/* Link with lowest tolerance determines timer interval */
+	if (keepalive_intv < n->keepalive_intv)
+		n->keepalive_intv = keepalive_intv;
+
+	/* Ensure link's abort limit corresponds to current interval */
+	l->abort_limit = l->tolerance / jiffies_to_msecs(n->keepalive_intv);
+}
+
 static void tipc_node_delete(struct tipc_node *node)
 {
 	list_del_rcu(&node->list);
@@ -160,8 +182,11 @@ void tipc_node_stop(struct net *net)
 	struct tipc_node *node, *t_node;
 
 	spin_lock_bh(&tn->node_list_lock);
-	list_for_each_entry_safe(node, t_node, &tn->node_list, list)
+	list_for_each_entry_safe(node, t_node, &tn->node_list, list) {
+		if (del_timer(&node->timer))
 			tipc_node_put(node);
+		tipc_node_put(node);
+	}
 	spin_unlock_bh(&tn->node_list_lock);
 }
 
@@ -219,131 +244,170 @@ void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port)
 	tipc_node_put(node);
 }
 
-/**
- * tipc_node_link_up - handle addition of link
- *
- * Link becomes active (alone or shared) or standby, depending on its priority.
+/* tipc_node_timeout - handle expiration of node timer
  */
-void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
+static void tipc_node_timeout(unsigned long data)
 {
-	struct tipc_link **active = &n_ptr->active_links[0];
-
-	n_ptr->working_links++;
-	n_ptr->action_flags |= TIPC_NOTIFY_LINK_UP;
-	n_ptr->link_id = l_ptr->peer_bearer_id << 16 | l_ptr->bearer_id;
-
-	pr_debug("Established link <%s> on network plane %c\n",
-		 l_ptr->name, l_ptr->net_plane);
-
-	if (!active[0]) {
-		active[0] = active[1] = l_ptr;
-		node_established_contact(n_ptr);
-		goto exit;
+	struct tipc_node *n = (struct tipc_node *)data;
+	struct sk_buff_head xmitq;
+	struct tipc_link *l;
+	struct tipc_media_addr *maddr;
+	int bearer_id;
+	int rc = 0;
+
+	__skb_queue_head_init(&xmitq);
+
+	for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) {
+		tipc_node_lock(n);
+		l = n->links[bearer_id].link;
+		if (l) {
+			/* Link tolerance may change asynchronously: */
+			tipc_node_calculate_timer(n, l);
+			rc = tipc_link_timeout(l, &xmitq);
+			if (rc & TIPC_LINK_DOWN_EVT)
+				tipc_link_reset(l);
 		}
-	if (l_ptr->priority < active[0]->priority) {
-		pr_debug("New link <%s> becomes standby\n", l_ptr->name);
-		goto exit;
+		tipc_node_unlock(n);
+		maddr = &n->links[bearer_id].maddr;
+		tipc_bearer_xmit(n->net, bearer_id, &xmitq, maddr);
 	}
-	tipc_link_dup_queue_xmit(active[0], l_ptr);
-	if (l_ptr->priority == active[0]->priority) {
-		active[0] = l_ptr;
-		goto exit;
-	}
-	pr_debug("Old link <%s> becomes standby\n", active[0]->name);
-	if (active[1] != active[0])
-		pr_debug("Old link <%s> becomes standby\n", active[1]->name);
-	active[0] = active[1] = l_ptr;
-exit:
-	/* Leave room for changeover header when returning 'mtu' to users: */
-	n_ptr->act_mtus[0] = active[0]->mtu - INT_H_SIZE;
-	n_ptr->act_mtus[1] = active[1]->mtu - INT_H_SIZE;
+	if (!mod_timer(&n->timer, jiffies + n->keepalive_intv))
+		tipc_node_get(n);
+	tipc_node_put(n);
 }
 
 /**
- * node_select_active_links - select active link
+ * tipc_node_link_up - handle addition of link
+ *
+ * Link becomes active (alone or shared) or standby, depending on its priority.
  */
-static void node_select_active_links(struct tipc_node *n_ptr)
+void tipc_node_link_up(struct tipc_node *n, int bearer_id)
 {
-	struct tipc_link **active = &n_ptr->active_links[0];
-	u32 i;
-	u32 highest_prio = 0;
+	int *slot0 = &n->active_links[0];
+	int *slot1 = &n->active_links[1];
+	struct tipc_link_entry *links = n->links;
+	struct tipc_link *l = n->links[bearer_id].link;
 
-	active[0] = active[1] = NULL;
+	/* Leave room for tunnel header when returning 'mtu' to users: */
+	links[bearer_id].mtu = l->mtu - INT_H_SIZE;
 
-	for (i = 0; i < MAX_BEARERS; i++) {
-		struct tipc_link *l_ptr = n_ptr->links[i];
+	n->working_links++;
+	n->action_flags |= TIPC_NOTIFY_LINK_UP;
+	n->link_id = l->peer_bearer_id << 16 | l->bearer_id;
 
-		if (!l_ptr || !tipc_link_is_up(l_ptr) ||
-		    (l_ptr->priority < highest_prio))
-			continue;
+	pr_debug("Established link <%s> on network plane %c\n",
+		 l->name, l->net_plane);
+
+	/* No active links ? => take both active slots */
+	if (*slot0 < 0) {
+		*slot0 = bearer_id;
+		*slot1 = bearer_id;
+		node_established_contact(n);
+		return;
+	}
 
-		if (l_ptr->priority > highest_prio) {
-			highest_prio = l_ptr->priority;
-			active[0] = active[1] = l_ptr;
-		} else {
-			active[1] = l_ptr;
+	/* Lower prio than current active ? => no slot */
+	if (l->priority < links[*slot0].link->priority) {
+		pr_debug("New link <%s> becomes standby\n", l->name);
+		return;
 	}
+	tipc_link_dup_queue_xmit(links[*slot0].link, l);
+
+	/* Same prio as current active ? => take one slot */
+	if (l->priority == links[*slot0].link->priority) {
+		*slot0 = bearer_id;
+		return;
 	}
+
+	/* Higher prio than current active => take both active slots */
+	pr_debug("Old link <%s> now standby\n", links[*slot0].link->name);
+	*slot0 = bearer_id;
+	*slot1 = bearer_id;
 }
 
 /**
  * tipc_node_link_down - handle loss of link
  */
-void tipc_node_link_down(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
+void tipc_node_link_down(struct tipc_node *n, int bearer_id)
 {
-	struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id);
-	struct tipc_link **active;
+	int *slot0 = &n->active_links[0];
+	int *slot1 = &n->active_links[1];
+	int i, highest = 0;
+	struct tipc_link *l, *_l;
 
-	n_ptr->working_links--;
-	n_ptr->action_flags |= TIPC_NOTIFY_LINK_DOWN;
-	n_ptr->link_id = l_ptr->peer_bearer_id << 16 | l_ptr->bearer_id;
+	l = n->links[bearer_id].link;
+	n->working_links--;
+	n->action_flags |= TIPC_NOTIFY_LINK_DOWN;
+	n->link_id = l->peer_bearer_id << 16 | l->bearer_id;
 
-	if (!tipc_link_is_active(l_ptr)) {
-		pr_debug("Lost standby link <%s> on network plane %c\n",
-			 l_ptr->name, l_ptr->net_plane);
-		return;
-	}
 	pr_debug("Lost link <%s> on network plane %c\n",
-		 l_ptr->name, l_ptr->net_plane);
-
-	active = &n_ptr->active_links[0];
-	if (active[0] == l_ptr)
-		active[0] = active[1];
-	if (active[1] == l_ptr)
-		active[1] = active[0];
-	if (active[0] == l_ptr)
-		node_select_active_links(n_ptr);
-	if (tipc_node_is_up(n_ptr))
-		tipc_link_failover_send_queue(l_ptr);
-	else
-		node_lost_contact(n_ptr);
+		 l->name, l->net_plane);
 
-	/* Leave room for changeover header when returning 'mtu' to users: */
-	if (active[0]) {
-		n_ptr->act_mtus[0] = active[0]->mtu - INT_H_SIZE;
-		n_ptr->act_mtus[1] = active[1]->mtu - INT_H_SIZE;
-		return;
+	/* Select new active link if any available */
+	*slot0 = INVALID_BEARER_ID;
+	*slot1 = INVALID_BEARER_ID;
+	for (i = 0; i < MAX_BEARERS; i++) {
+		_l = n->links[i].link;
+		if (!_l || !tipc_link_is_up(_l))
+			continue;
+		if (_l->priority < highest)
+			continue;
+		if (_l->priority > highest) {
+			highest = _l->priority;
+			*slot0 = i;
+			*slot1 = i;
+			continue;
 		}
-	/* Loopback link went down? No fragmentation needed from now on. */
-	if (n_ptr->addr == tn->own_addr) {
-		n_ptr->act_mtus[0] = MAX_MSG_SIZE;
-		n_ptr->act_mtus[1] = MAX_MSG_SIZE;
+		*slot1 = i;
 	}
+	if (tipc_node_is_up(n))
+		tipc_link_failover_send_queue(l);
+	else
+		node_lost_contact(n);
+}
+
+bool tipc_node_is_up(struct tipc_node *n)
+{
+	return n->active_links[0] != INVALID_BEARER_ID;
 }
 
-int tipc_node_active_links(struct tipc_node *n_ptr)
+void tipc_node_check_dest(struct tipc_node *n, struct tipc_bearer *b,
+			  bool *link_up, bool *addr_match,
+			  struct tipc_media_addr *maddr)
 {
-	return n_ptr->active_links[0] != NULL;
+	struct tipc_link *l = n->links[b->identity].link;
+	struct tipc_media_addr *curr = &n->links[b->identity].maddr;
+
+	*link_up = l && tipc_link_is_up(l);
+	*addr_match = l && !memcmp(curr, maddr, sizeof(*maddr));
 }
 
-int tipc_node_is_up(struct tipc_node *n_ptr)
+bool tipc_node_update_dest(struct tipc_node *n,  struct tipc_bearer *b,
+			   struct tipc_media_addr *maddr)
 {
-	return tipc_node_active_links(n_ptr);
+	struct tipc_link *l = n->links[b->identity].link;
+	struct tipc_media_addr *curr = &n->links[b->identity].maddr;
+	struct sk_buff_head *inputq = &n->links[b->identity].inputq;
+
+	if (!l) {
+		l = tipc_link_create(n, b, maddr, inputq, &n->bclink.namedq);
+		if (!l)
+			return false;
+		tipc_node_calculate_timer(n, l);
+		if (n->link_cnt == 1) {
+			if (!mod_timer(&n->timer, jiffies + n->keepalive_intv))
+				tipc_node_get(n);
+		}
+	}
+	memcpy(&l->media_addr, maddr, sizeof(*maddr));
+	memcpy(curr, maddr, sizeof(*maddr));
+	tipc_link_reset(l);
+	return true;
 }
 
 void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
 {
-	n_ptr->links[l_ptr->bearer_id] = l_ptr;
+	n_ptr->links[l_ptr->bearer_id].link = l_ptr;
 	n_ptr->link_cnt++;
 }
 
@@ -352,15 +416,151 @@ void tipc_node_detach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
 	int i;
 
 	for (i = 0; i < MAX_BEARERS; i++) {
-		if (l_ptr != n_ptr->links[i])
+		if (l_ptr != n_ptr->links[i].link)
 			continue;
-		n_ptr->links[i] = NULL;
+		n_ptr->links[i].link = NULL;
 		n_ptr->link_cnt--;
 	}
 }
 
+/* tipc_node_fsm_evt - node finite state machine
+ * Determines when contact is allowed with peer node
+ */
+static void tipc_node_fsm_evt(struct tipc_node *n, int evt)
+{
+	int state = n->state;
+
+	switch (state) {
+	case SELF_DOWN_PEER_DOWN:
+		switch (evt) {
+		case SELF_ESTABL_CONTACT_EVT:
+			state = SELF_UP_PEER_COMING;
+			break;
+		case PEER_ESTABL_CONTACT_EVT:
+			state = SELF_COMING_PEER_UP;
+			break;
+		case SELF_LOST_CONTACT_EVT:
+		case PEER_LOST_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	case SELF_UP_PEER_UP:
+		switch (evt) {
+		case SELF_LOST_CONTACT_EVT:
+			state = SELF_DOWN_PEER_LEAVING;
+			break;
+		case PEER_LOST_CONTACT_EVT:
+			state = SELF_LEAVING_PEER_DOWN;
+			break;
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	case SELF_DOWN_PEER_LEAVING:
+		switch (evt) {
+		case PEER_LOST_CONTACT_EVT:
+			state = SELF_DOWN_PEER_DOWN;
+			break;
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+		case SELF_LOST_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	case SELF_UP_PEER_COMING:
+		switch (evt) {
+		case PEER_ESTABL_CONTACT_EVT:
+			state = SELF_UP_PEER_UP;
+			break;
+		case SELF_LOST_CONTACT_EVT:
+			state = SELF_DOWN_PEER_LEAVING;
+			break;
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_LOST_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	case SELF_COMING_PEER_UP:
+		switch (evt) {
+		case SELF_ESTABL_CONTACT_EVT:
+			state = SELF_UP_PEER_UP;
+			break;
+		case PEER_LOST_CONTACT_EVT:
+			state = SELF_LEAVING_PEER_DOWN;
+			break;
+		case SELF_LOST_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	case SELF_LEAVING_PEER_DOWN:
+		switch (evt) {
+		case SELF_LOST_CONTACT_EVT:
+			state = SELF_DOWN_PEER_DOWN;
+			break;
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+		case PEER_LOST_CONTACT_EVT:
+			break;
+		default:
+			pr_err("Unknown node fsm evt %x/%x\n", state, evt);
+		}
+		break;
+	default:
+		pr_err("Unknown node fsm state %x\n", state);
+		break;
+	}
+
+	n->state = state;
+}
+
+bool tipc_node_filter_skb(struct tipc_node *n, struct tipc_link *l,
+			  struct tipc_msg *hdr)
+{
+	int state = n->state;
+
+	if (likely(state == SELF_UP_PEER_UP))
+		return true;
+
+	if (state == SELF_DOWN_PEER_DOWN)
+		return true;
+
+	if (state == SELF_UP_PEER_COMING) {
+		/* If not traffic msg, peer may still be ESTABLISHING */
+		if (tipc_link_is_up(l) && msg_is_traffic(hdr))
+			tipc_node_fsm_evt(n, PEER_ESTABL_CONTACT_EVT);
+		return true;
+	}
+
+	if (state == SELF_COMING_PEER_UP)
+		return true;
+
+	if (state == SELF_LEAVING_PEER_DOWN)
+		return false;
+
+	if (state == SELF_DOWN_PEER_LEAVING) {
+		if (msg_peer_is_up(hdr))
+			return false;
+		tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT);
+		return true;
+	}
+	return false;
+}
+
 static void node_established_contact(struct tipc_node *n_ptr)
 {
+	tipc_node_fsm_evt(n_ptr, SELF_ESTABL_CONTACT_EVT);
 	n_ptr->action_flags |= TIPC_NOTIFY_NODE_UP;
 	n_ptr->bclink.oos_state = 0;
 	n_ptr->bclink.acked = tipc_bclink_get_last_sent(n_ptr->net);
@@ -396,21 +596,18 @@ static void node_lost_contact(struct tipc_node *n_ptr)
 
 	/* Abort any ongoing link failover */
 	for (i = 0; i < MAX_BEARERS; i++) {
-		struct tipc_link *l_ptr = n_ptr->links[i];
+		struct tipc_link *l_ptr = n_ptr->links[i].link;
 		if (!l_ptr)
 			continue;
-		l_ptr->flags &= ~LINK_FAILINGOVER;
+		l_ptr->exec_mode = TIPC_LINK_OPEN;
 		l_ptr->failover_checkpt = 0;
 		l_ptr->failover_pkts = 0;
 		kfree_skb(l_ptr->failover_skb);
 		l_ptr->failover_skb = NULL;
 		tipc_link_reset_fragments(l_ptr);
 	}
-
-	n_ptr->action_flags &= ~TIPC_WAIT_OWN_LINKS_DOWN;
-
 	/* Prevent re-contact with node until cleanup is done */
-	n_ptr->action_flags |= TIPC_WAIT_PEER_LINKS_DOWN;
+	tipc_node_fsm_evt(n_ptr, SELF_LOST_CONTACT_EVT);
 
 	/* Notify publications from this node */
 	n_ptr->action_flags |= TIPC_NOTIFY_NODE_DOWN;
@@ -453,7 +650,7 @@ int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 addr,
 		goto exit;
 
 	tipc_node_lock(node);
-	link = node->links[bearer_id];
+	link = node->links[bearer_id].link;
 	if (link) {
 		strncpy(linkname, link->name, len);
 		err = 0;
@@ -559,6 +756,160 @@ static int __tipc_nl_add_node(struct tipc_nl_msg *msg, struct tipc_node *node)
 	return -EMSGSIZE;
 }
 
+static struct tipc_link *tipc_node_select_link(struct tipc_node *n, int sel,
+					       int *bearer_id,
+					       struct tipc_media_addr **maddr)
+{
+	int id = n->active_links[sel & 1];
+
+	if (unlikely(id < 0))
+		return NULL;
+
+	*bearer_id = id;
+	*maddr = &n->links[id].maddr;
+	return n->links[id].link;
+}
+
+/**
+ * tipc_node_xmit() is the general link level function for message sending
+ * @net: the applicable net namespace
+ * @list: chain of buffers containing message
+ * @dnode: address of destination node
+ * @selector: a number used for deterministic link selection
+ * Consumes the buffer chain, except when returning -ELINKCONG
+ * Returns 0 if success, otherwise errno: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE
+ */
+int tipc_node_xmit(struct net *net, struct sk_buff_head *list,
+		   u32 dnode, int selector)
+{
+	struct tipc_link *l = NULL;
+	struct tipc_node *n;
+	struct sk_buff_head xmitq;
+	struct tipc_media_addr *maddr;
+	int bearer_id;
+	int rc = -EHOSTUNREACH;
+
+	__skb_queue_head_init(&xmitq);
+	n = tipc_node_find(net, dnode);
+	if (likely(n)) {
+		tipc_node_lock(n);
+		l = tipc_node_select_link(n, selector, &bearer_id, &maddr);
+		if (likely(l))
+			rc = tipc_link_xmit(l, list, &xmitq);
+		if (unlikely(rc == -ENOBUFS))
+			tipc_link_reset(l);
+		tipc_node_unlock(n);
+		tipc_node_put(n);
+	}
+	if (likely(!rc)) {
+		tipc_bearer_xmit(net, bearer_id, &xmitq, maddr);
+		return 0;
+	}
+	if (likely(in_own_node(net, dnode))) {
+		tipc_sk_rcv(net, list);
+		return 0;
+	}
+	return rc;
+}
+
+/* tipc_node_xmit_skb(): send single buffer to destination
+ * Buffers sent via this functon are generally TIPC_SYSTEM_IMPORTANCE
+ * messages, which will not be rejected
+ * The only exception is datagram messages rerouted after secondary
+ * lookup, which are rare and safe to dispose of anyway.
+ * TODO: Return real return value, and let callers use
+ * tipc_wait_for_sendpkt() where applicable
+ */
+int tipc_node_xmit_skb(struct net *net, struct sk_buff *skb, u32 dnode,
+		       u32 selector)
+{
+	struct sk_buff_head head;
+	int rc;
+
+	skb_queue_head_init(&head);
+	__skb_queue_tail(&head, skb);
+	rc = tipc_node_xmit(net, &head, dnode, selector);
+	if (rc == -ELINKCONG)
+		kfree_skb(skb);
+	return 0;
+}
+
+/**
+ * tipc_rcv - process TIPC packets/messages arriving from off-node
+ * @net: the applicable net namespace
+ * @skb: TIPC packet
+ * @bearer: pointer to bearer message arrived on
+ *
+ * Invoked with no locks held. Bearer pointer must point to a valid bearer
+ * structure (i.e. cannot be NULL), but bearer can be inactive.
+ */
+void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b)
+{
+	struct sk_buff_head xmitq;
+	struct tipc_node *n;
+	struct tipc_link *l;
+	struct tipc_msg *hdr;
+	struct tipc_media_addr *maddr;
+	int bearer_id = b->identity;
+	int rc = 0;
+
+	__skb_queue_head_init(&xmitq);
+
+	/* Ensure message is well-formed */
+	if (unlikely(!tipc_msg_validate(skb)))
+		goto discard;
+
+	/* Handle arrival of a non-unicast link packet */
+	hdr = buf_msg(skb);
+	if (unlikely(msg_non_seq(hdr))) {
+		if (msg_user(hdr) ==  LINK_CONFIG)
+			tipc_disc_rcv(net, skb, b);
+		else
+			tipc_bclink_rcv(net, skb);
+		return;
+	}
+
+	/* Locate neighboring node that sent packet */
+	n = tipc_node_find(net, msg_prevnode(hdr));
+	if (unlikely(!n))
+		goto discard;
+	tipc_node_lock(n);
+
+	/* Locate link endpoint that should handle packet */
+	l = n->links[bearer_id].link;
+	if (unlikely(!l))
+		goto unlock;
+
+	/* Is reception of this packet permitted at the moment ? */
+	if (unlikely(n->state != SELF_UP_PEER_UP))
+		if (!tipc_node_filter_skb(n, l, hdr))
+			goto unlock;
+
+	if (unlikely(msg_user(hdr) == LINK_PROTOCOL))
+		tipc_bclink_sync_state(n, hdr);
+
+	/* Release acked broadcast messages */
+	if (unlikely(n->bclink.acked != msg_bcast_ack(hdr)))
+		tipc_bclink_acknowledge(n, msg_bcast_ack(hdr));
+
+	/* Check protocol and update link state */
+	rc = tipc_link_rcv(l, skb, &xmitq);
+
+	if (unlikely(rc & TIPC_LINK_UP_EVT))
+		tipc_link_activate(l);
+	if (unlikely(rc & TIPC_LINK_DOWN_EVT))
+		tipc_link_reset(l);
+	skb = NULL;
+unlock:
+	tipc_node_unlock(n);
+	tipc_sk_rcv(net, &n->links[bearer_id].inputq);
+	maddr = &n->links[bearer_id].maddr;
+	tipc_bearer_xmit(net, bearer_id, &xmitq, maddr);
+	tipc_node_put(n);
+discard:
+	kfree_skb(skb);
+}
+
 int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	int err;

net/tipc/node.h

@@ -45,6 +45,26 @@
 /* Out-of-range value for node signature */
 #define INVALID_NODE_SIG	0x10000
 
+#define INVALID_BEARER_ID -1
+
+/* Node FSM states and events:
+ */
+enum {
+	SELF_DOWN_PEER_DOWN    = 0xdd,
+	SELF_UP_PEER_UP        = 0xaa,
+	SELF_DOWN_PEER_LEAVING = 0xd1,
+	SELF_UP_PEER_COMING    = 0xac,
+	SELF_COMING_PEER_UP    = 0xca,
+	SELF_LEAVING_PEER_DOWN = 0x1d,
+};
+
+enum {
+	SELF_ESTABL_CONTACT_EVT = 0xec,
+	SELF_LOST_CONTACT_EVT   = 0x1c,
+	PEER_ESTABL_CONTACT_EVT = 0xfec,
+	PEER_LOST_CONTACT_EVT   = 0xf1c
+};
+
 /* Flags used to take different actions according to flag type
  * TIPC_WAIT_PEER_LINKS_DOWN: wait to see that peer's links are down
  * TIPC_WAIT_OWN_LINKS_DOWN: wait until peer node is declared down
@@ -54,8 +74,6 @@
  */
 enum {
 	TIPC_MSG_EVT                    = 1,
-	TIPC_WAIT_PEER_LINKS_DOWN	= (1 << 1),
-	TIPC_WAIT_OWN_LINKS_DOWN	= (1 << 2),
 	TIPC_NOTIFY_NODE_DOWN		= (1 << 3),
 	TIPC_NOTIFY_NODE_UP		= (1 << 4),
 	TIPC_WAKEUP_BCAST_USERS		= (1 << 5),
@@ -85,10 +103,17 @@ struct tipc_node_bclink {
 	u32 deferred_size;
 	struct sk_buff_head deferdq;
 	struct sk_buff *reasm_buf;
-	int inputq_map;
+	struct sk_buff_head namedq;
 	bool recv_permitted;
 };
 
+struct tipc_link_entry {
+	struct tipc_link *link;
+	u32 mtu;
+	struct sk_buff_head inputq;
+	struct tipc_media_addr maddr;
+};
+
 /**
  * struct tipc_node - TIPC node structure
  * @addr: network address of node
@@ -98,9 +123,8 @@ struct tipc_node_bclink {
  * @hash: links to adjacent nodes in unsorted hash chain
  * @inputq: pointer to input queue containing messages for msg event
  * @namedq: pointer to name table input queue with name table messages
- * @curr_link: the link holding the node lock, if any
- * @active_links: pointers to active links to node
- * @links: pointers to all links to node
+ * @active_links: bearer ids of active links, used as index into links[] array
+ * @links: array containing references to all links to node
  * @action_flags: bit mask of different types of node actions
  * @bclink: broadcast-related info
  * @list: links to adjacent nodes in sorted list of cluster's nodes
@@ -120,12 +144,12 @@ struct tipc_node {
 	struct hlist_node hash;
 	struct sk_buff_head *inputq;
 	struct sk_buff_head *namedq;
-	struct tipc_link *active_links[2];
-	u32 act_mtus[2];
-	struct tipc_link *links[MAX_BEARERS];
+	int active_links[2];
+	struct tipc_link_entry links[MAX_BEARERS];
 	int action_flags;
 	struct tipc_node_bclink bclink;
 	struct list_head list;
+	int state;
 	int link_cnt;
 	u16 working_links;
 	u16 capabilities;
@@ -133,6 +157,8 @@ struct tipc_node {
 	u32 link_id;
 	struct list_head publ_list;
 	struct list_head conn_sks;
+	unsigned long keepalive_intv;
+	struct timer_list timer;
 	struct rcu_head rcu;
 };
 
@@ -140,18 +166,25 @@ struct tipc_node *tipc_node_find(struct net *net, u32 addr);
 void tipc_node_put(struct tipc_node *node);
 struct tipc_node *tipc_node_create(struct net *net, u32 addr);
 void tipc_node_stop(struct net *net);
+void tipc_node_check_dest(struct tipc_node *n, struct tipc_bearer *bearer,
+			  bool *link_up, bool *addr_match,
+			  struct tipc_media_addr *maddr);
+bool tipc_node_update_dest(struct tipc_node *n, struct tipc_bearer *bearer,
+			   struct tipc_media_addr *maddr);
 void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr);
 void tipc_node_detach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr);
-void tipc_node_link_down(struct tipc_node *n_ptr, struct tipc_link *l_ptr);
-void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr);
-int tipc_node_active_links(struct tipc_node *n_ptr);
-int tipc_node_is_up(struct tipc_node *n_ptr);
+void tipc_node_link_down(struct tipc_node *n_ptr, int bearer_id);
+void tipc_node_link_up(struct tipc_node *n_ptr, int bearer_id);
+bool tipc_node_is_up(struct tipc_node *n);
 int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 node,
 			   char *linkname, size_t len);
 void tipc_node_unlock(struct tipc_node *node);
+int tipc_node_xmit(struct net *net, struct sk_buff_head *list, u32 dnode,
+		   int selector);
+int tipc_node_xmit_skb(struct net *net, struct sk_buff *skb, u32 dest,
+		       u32 selector);
 int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port);
 void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port);
-
 int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb);
 
 static inline void tipc_node_lock(struct tipc_node *node)
@@ -159,26 +192,30 @@ static inline void tipc_node_lock(struct tipc_node *node)
 	spin_lock_bh(&node->lock);
 }
 
-static inline bool tipc_node_blocked(struct tipc_node *node)
+static inline struct tipc_link *node_active_link(struct tipc_node *n, int sel)
 {
-	return (node->action_flags & (TIPC_WAIT_PEER_LINKS_DOWN |
-		TIPC_NOTIFY_NODE_DOWN | TIPC_WAIT_OWN_LINKS_DOWN));
+	int bearer_id = n->active_links[sel & 1];
+
+	if (unlikely(bearer_id == INVALID_BEARER_ID))
+		return NULL;
+
+	return n->links[bearer_id].link;
 }
 
-static inline uint tipc_node_get_mtu(struct net *net, u32 addr, u32 selector)
+static inline unsigned int tipc_node_get_mtu(struct net *net, u32 addr, u32 sel)
 {
-	struct tipc_node *node;
-	u32 mtu;
+	struct tipc_node *n;
+	int bearer_id;
+	unsigned int mtu = MAX_MSG_SIZE;
 
-	node = tipc_node_find(net, addr);
-
-	if (likely(node)) {
-		mtu = node->act_mtus[selector & 1];
-		tipc_node_put(node);
-	} else {
-		mtu = MAX_MSG_SIZE;
-	}
+	n = tipc_node_find(net, addr);
+	if (unlikely(!n))
+		return mtu;
 
+	bearer_id = n->active_links[sel & 1];
+	if (likely(bearer_id != INVALID_BEARER_ID))
+		mtu = n->links[bearer_id].mtu;
+	tipc_node_put(n);
 	return mtu;
 }
 

net/tipc/socket.c

@@ -261,7 +261,7 @@ static void tsk_rej_rx_queue(struct sock *sk)
 
 	while ((skb = __skb_dequeue(&sk->sk_receive_queue))) {
 		if (tipc_msg_reverse(own_node, skb, &dnode, TIPC_ERR_NO_PORT))
-			tipc_link_xmit_skb(sock_net(sk), skb, dnode, 0);
+			tipc_node_xmit_skb(sock_net(sk), skb, dnode, 0);
 	}
 }
 
@@ -443,7 +443,7 @@ static int tipc_release(struct socket *sock)
 			}
 			if (tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode,
 					     TIPC_ERR_NO_PORT))
-				tipc_link_xmit_skb(net, skb, dnode, 0);
+				tipc_node_xmit_skb(net, skb, dnode, 0);
 		}
 	}
 
@@ -456,7 +456,7 @@ static int tipc_release(struct socket *sock)
 				      tsk_own_node(tsk), tsk_peer_port(tsk),
 				      tsk->portid, TIPC_ERR_NO_PORT);
 		if (skb)
-			tipc_link_xmit_skb(net, skb, dnode, tsk->portid);
+			tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
 		tipc_node_remove_conn(net, dnode, tsk->portid);
 	}
 
@@ -686,21 +686,22 @@ static int tipc_sendmcast(struct  socket *sock, struct tipc_name_seq *seq,
 
 	do {
 		rc = tipc_bclink_xmit(net, pktchain);
-		if (likely(rc >= 0)) {
-			rc = dsz;
-			break;
+		if (likely(!rc))
+			return dsz;
+
+		if (rc == -ELINKCONG) {
+			tsk->link_cong = 1;
+			rc = tipc_wait_for_sndmsg(sock, &timeo);
+			if (!rc)
+				continue;
 		}
+		__skb_queue_purge(pktchain);
 		if (rc == -EMSGSIZE) {
 			msg->msg_iter = save;
 			goto new_mtu;
 		}
-		if (rc != -ELINKCONG)
 		break;
-		tipc_sk(sk)->link_cong = 1;
-		rc = tipc_wait_for_sndmsg(sock, &timeo);
-		if (rc)
-			__skb_queue_purge(pktchain);
-	} while (!rc);
+	} while (1);
 	return rc;
 }
 
@@ -924,24 +925,25 @@ static int __tipc_sendmsg(struct socket *sock, struct msghdr *m, size_t dsz)
 	do {
 		skb = skb_peek(pktchain);
 		TIPC_SKB_CB(skb)->wakeup_pending = tsk->link_cong;
-		rc = tipc_link_xmit(net, pktchain, dnode, tsk->portid);
-		if (likely(rc >= 0)) {
+		rc = tipc_node_xmit(net, pktchain, dnode, tsk->portid);
+		if (likely(!rc)) {
 			if (sock->state != SS_READY)
 				sock->state = SS_CONNECTING;
-			rc = dsz;
-			break;
+			return dsz;
 		}
+		if (rc == -ELINKCONG) {
+			tsk->link_cong = 1;
+			rc = tipc_wait_for_sndmsg(sock, &timeo);
+			if (!rc)
+				continue;
+		}
+		__skb_queue_purge(pktchain);
 		if (rc == -EMSGSIZE) {
 			m->msg_iter = save;
 			goto new_mtu;
 		}
-		if (rc != -ELINKCONG)
 		break;
-		tsk->link_cong = 1;
-		rc = tipc_wait_for_sndmsg(sock, &timeo);
-		if (rc)
-			__skb_queue_purge(pktchain);
-	} while (!rc);
+	} while (1);
 
 	return rc;
 }
@@ -1043,15 +1045,16 @@ static int __tipc_send_stream(struct socket *sock, struct msghdr *m, size_t dsz)
 		return rc;
 	do {
 		if (likely(!tsk_conn_cong(tsk))) {
-			rc = tipc_link_xmit(net, pktchain, dnode, portid);
+			rc = tipc_node_xmit(net, pktchain, dnode, portid);
 			if (likely(!rc)) {
 				tsk->sent_unacked++;
 				sent += send;
 				if (sent == dsz)
-					break;
+					return dsz;
 				goto next;
 			}
 			if (rc == -EMSGSIZE) {
+				__skb_queue_purge(pktchain);
 				tsk->max_pkt = tipc_node_get_mtu(net, dnode,
 								 portid);
 				m->msg_iter = save;
@@ -1059,13 +1062,13 @@ static int __tipc_send_stream(struct socket *sock, struct msghdr *m, size_t dsz)
 			}
 			if (rc != -ELINKCONG)
 				break;
+
 			tsk->link_cong = 1;
 		}
 		rc = tipc_wait_for_sndpkt(sock, &timeo);
-		if (rc)
-			__skb_queue_purge(pktchain);
 	} while (!rc);
 
+	__skb_queue_purge(pktchain);
 	return sent ? sent : rc;
 }
 
@@ -1221,7 +1224,7 @@ static void tipc_sk_send_ack(struct tipc_sock *tsk, uint ack)
 		return;
 	msg = buf_msg(skb);
 	msg_set_msgcnt(msg, ack);
-	tipc_link_xmit_skb(net, skb, dnode, msg_link_selector(msg));
+	tipc_node_xmit_skb(net, skb, dnode, msg_link_selector(msg));
 }
 
 static int tipc_wait_for_rcvmsg(struct socket *sock, long *timeop)
@@ -1700,7 +1703,7 @@ static int tipc_backlog_rcv(struct sock *sk, struct sk_buff *skb)
 		return 0;
 	}
 	if (!err || tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode, -err))
-		tipc_link_xmit_skb(net, skb, dnode, tsk->portid);
+		tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
 	return 0;
 }
 
@@ -1796,7 +1799,7 @@ int tipc_sk_rcv(struct net *net, struct sk_buff_head *inputq)
 		if (!tipc_msg_reverse(tn->own_addr, skb, &dnode, -err))
 			continue;
 xmit:
-		tipc_link_xmit_skb(net, skb, dnode, dport);
+		tipc_node_xmit_skb(net, skb, dnode, dport);
 	}
 	return err ? -EHOSTUNREACH : 0;
 }
@@ -2089,7 +2092,7 @@ static int tipc_shutdown(struct socket *sock, int how)
 			}
 			if (tipc_msg_reverse(tsk_own_node(tsk), skb, &dnode,
 					     TIPC_CONN_SHUTDOWN))
-				tipc_link_xmit_skb(net, skb, dnode,
+				tipc_node_xmit_skb(net, skb, dnode,
 						   tsk->portid);
 		} else {
 			dnode = tsk_peer_node(tsk);
@@ -2099,7 +2102,7 @@ static int tipc_shutdown(struct socket *sock, int how)
 					      0, dnode, tsk_own_node(tsk),
 					      tsk_peer_port(tsk),
 					      tsk->portid, TIPC_CONN_SHUTDOWN);
-			tipc_link_xmit_skb(net, skb, dnode, tsk->portid);
+			tipc_node_xmit_skb(net, skb, dnode, tsk->portid);
 		}
 		tsk->connected = 0;
 		sock->state = SS_DISCONNECTING;
@@ -2161,7 +2164,7 @@ static void tipc_sk_timeout(unsigned long data)
 	}
 	bh_unlock_sock(sk);
 	if (skb)
-		tipc_link_xmit_skb(sock_net(sk), skb, peer_node, tsk->portid);
+		tipc_node_xmit_skb(sock_net(sk), skb, peer_node, tsk->portid);
 exit:
 	sock_put(sk);
 }