Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking updates from David Miller:
 "Another merge window, another set of networking changes.  I've heard
  rumblings that the lightweight tunnels infrastructure has been voted
  networking change of the year.  But what do I know?

   1) Add conntrack support to openvswitch, from Joe Stringer.

   2) Initial support for VRF (Virtual Routing and Forwarding), which
      allows the segmentation of routing paths without using multiple
      devices.  There are some semantic kinks to work out still, but
      this is a reasonably strong foundation.  From David Ahern.

   3) Remove spinlock fro act_bpf fast path, from Alexei Starovoitov.

   4) Ignore route nexthops with a link down state in ipv6, just like
      ipv4.  From Andy Gospodarek.

   5) Remove spinlock from fast path of act_gact and act_mirred, from
      Eric Dumazet.

   6) Document the DSA layer, from Florian Fainelli.

   7) Add netconsole support to bcmgenet, systemport, and DSA.  Also
      from Florian Fainelli.

   8) Add Mellanox Switch Driver and core infrastructure, from Jiri
      Pirko.

   9) Add support for "light weight tunnels", which allow for
      encapsulation and decapsulation without bearing the overhead of a
      full blown netdevice.  From Thomas Graf, Jiri Benc, and a cast of
      others.

  10) Add Identifier Locator Addressing support for ipv6, from Tom
      Herbert.

  11) Support fragmented SKBs in iwlwifi, from Johannes Berg.

  12) Allow perf PMUs to be accessed from eBPF programs, from Kaixu Xia.

  13) Add BQL support to 3c59x driver, from Loganaden Velvindron.

  14) Stop using a zero TX queue length to mean that a device shouldn't
      have a qdisc attached, use an explicit flag instead.  From Phil
      Sutter.

  15) Use generic geneve netdevice infrastructure in openvswitch, from
      Pravin B Shelar.

  16) Add infrastructure to avoid re-forwarding a packet in software
      that was already forwarded by a hardware switch.  From Scott
      Feldman.

  17) Allow AF_PACKET fanout function to be implemented in a bpf
      program, from Willem de Bruijn"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1458 commits)
  netfilter: nf_conntrack: make nf_ct_zone_dflt built-in
  netfilter: nf_dup{4, 6}: fix build error when nf_conntrack disabled
  net: fec: clear receive interrupts before processing a packet
  ipv6: fix exthdrs offload registration in out_rt path
  xen-netback: add support for multicast control
  bgmac: Update fixed_phy_register()
  sock, diag: fix panic in sock_diag_put_filterinfo
  flow_dissector: Use 'const' where possible.
  flow_dissector: Fix function argument ordering dependency
  ixgbe: Resolve "initialized field overwritten" warnings
  ixgbe: Remove bimodal SR-IOV disabling
  ixgbe: Add support for reporting 2.5G link speed
  ixgbe: fix bounds checking in ixgbe_setup_tc for 82598
  ixgbe: support for ethtool set_rxfh
  ixgbe: Avoid needless PHY access on copper phys
  ixgbe: cleanup to use cached mask value
  ixgbe: Remove second instance of lan_id variable
  ixgbe: use kzalloc for allocating one thing
  flow: Move __get_hash_from_flowi{4,6} into flow_dissector.c
  ixgbe: Remove unused PCI bus types
  ...

1 files changed, 844 insertions, 135 deletions

diff --git a/net/tipc/node.c b/net/tipc/node.c
index 0b1d61a..703875f 100644
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -40,10 +40,42 @@
 #include "name_distr.h"
 #include "socket.h"
 #include "bcast.h"
+#include "discover.h"
 
-static void node_lost_contact(struct tipc_node *n_ptr);
+/* 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,
+	NODE_FAILINGOVER       = 0xf0,
+	NODE_SYNCHING          = 0xcc
+};
+
+enum {
+	SELF_ESTABL_CONTACT_EVT = 0xece,
+	SELF_LOST_CONTACT_EVT   = 0x1ce,
+	PEER_ESTABL_CONTACT_EVT = 0x9ece,
+	PEER_LOST_CONTACT_EVT   = 0x91ce,
+	NODE_FAILOVER_BEGIN_EVT = 0xfbe,
+	NODE_FAILOVER_END_EVT   = 0xfee,
+	NODE_SYNCH_BEGIN_EVT    = 0xcbe,
+	NODE_SYNCH_END_EVT      = 0xcee
+};
+
+static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id,
+				  struct sk_buff_head *xmitq,
+				  struct tipc_media_addr **maddr);
+static void tipc_node_link_down(struct tipc_node *n, int bearer_id,
+				bool delete);
+static void node_lost_contact(struct tipc_node *n, struct sk_buff_head *inputq);
 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;
@@ -110,7 +142,7 @@ struct tipc_node *tipc_node_find(struct net *net, u32 addr)
 	return NULL;
 }
 
-struct tipc_node *tipc_node_create(struct net *net, u32 addr)
+struct tipc_node *tipc_node_create(struct net *net, u32 addr, u16 capabilities)
 {
 	struct tipc_net *tn = net_generic(net, tipc_net_id);
 	struct tipc_node *n_ptr, *temp_node;
@@ -126,12 +158,14 @@ struct tipc_node *tipc_node_create(struct net *net, u32 addr)
 	}
 	n_ptr->addr = addr;
 	n_ptr->net = net;
+	n_ptr->capabilities = capabilities;
 	kref_init(&n_ptr->kref);
 	spin_lock_init(&n_ptr->lock);
 	INIT_HLIST_NODE(&n_ptr->hash);
 	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 +173,32 @@ 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 +212,11 @@ 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);
 }
 
@@ -219,158 +274,551 @@ void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port)
 	tipc_node_put(node);
 }
 
+/* tipc_node_timeout - handle expiration of node timer
+ */
+static void tipc_node_timeout(unsigned long data)
+{
+	struct tipc_node *n = (struct tipc_node *)data;
+	struct tipc_link_entry *le;
+	struct sk_buff_head xmitq;
+	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);
+		le = &n->links[bearer_id];
+		if (le->link) {
+			/* Link tolerance may change asynchronously: */
+			tipc_node_calculate_timer(n, le->link);
+			rc = tipc_link_timeout(le->link, &xmitq);
+		}
+		tipc_node_unlock(n);
+		tipc_bearer_xmit(n->net, bearer_id, &xmitq, &le->maddr);
+		if (rc & TIPC_LINK_DOWN_EVT)
+			tipc_node_link_down(n, bearer_id, false);
+	}
+	if (!mod_timer(&n->timer, jiffies + n->keepalive_intv))
+		tipc_node_get(n);
+	tipc_node_put(n);
+}
+
 /**
- * tipc_node_link_up - handle addition of link
- *
+ * __tipc_node_link_up - handle addition of link
+ * Node lock must be held by caller
  * Link becomes active (alone or shared) or standby, depending on its priority.
  */
-void tipc_node_link_up(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
+static void __tipc_node_link_up(struct tipc_node *n, int bearer_id,
+				struct sk_buff_head *xmitq)
 {
-	struct tipc_link **active = &n_ptr->active_links[0];
+	int *slot0 = &n->active_links[0];
+	int *slot1 = &n->active_links[1];
+	struct tipc_link *ol = node_active_link(n, 0);
+	struct tipc_link *nl = n->links[bearer_id].link;
 
-	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;
+	if (!nl || !tipc_link_is_up(nl))
+		return;
 
-	pr_debug("Established link <%s> on network plane %c\n",
-		 l_ptr->name, l_ptr->net_plane);
+	n->working_links++;
+	n->action_flags |= TIPC_NOTIFY_LINK_UP;
+	n->link_id = nl->peer_bearer_id << 16 | bearer_id;
 
-	if (!active[0]) {
-		active[0] = active[1] = l_ptr;
-		node_established_contact(n_ptr);
-		goto exit;
-	}
-	if (l_ptr->priority < active[0]->priority) {
-		pr_debug("New link <%s> becomes standby\n", l_ptr->name);
-		goto exit;
+	/* Leave room for tunnel header when returning 'mtu' to users: */
+	n->links[bearer_id].mtu = nl->mtu - INT_H_SIZE;
+
+	tipc_bearer_add_dest(n->net, bearer_id, n->addr);
+
+	pr_debug("Established link <%s> on network plane %c\n",
+		 nl->name, nl->net_plane);
+
+	/* First link? => give it both slots */
+	if (!ol) {
+		*slot0 = bearer_id;
+		*slot1 = bearer_id;
+		tipc_link_build_bcast_sync_msg(nl, xmitq);
+		node_established_contact(n);
+		return;
 	}
-	tipc_link_dup_queue_xmit(active[0], l_ptr);
-	if (l_ptr->priority == active[0]->priority) {
-		active[0] = l_ptr;
-		goto exit;
+
+	/* Second link => redistribute slots */
+	if (nl->priority > ol->priority) {
+		pr_debug("Old link <%s> becomes standby\n", ol->name);
+		*slot0 = bearer_id;
+		*slot1 = bearer_id;
+	} else if (nl->priority == ol->priority) {
+		*slot0 = bearer_id;
+	} else {
+		pr_debug("New link <%s> is standby\n", nl->name);
 	}
-	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;
+
+	/* Prepare synchronization with first link */
+	tipc_link_tnl_prepare(ol, nl, SYNCH_MSG, xmitq);
 }
 
 /**
- * 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)
+static void tipc_node_link_up(struct tipc_node *n, int bearer_id,
+			      struct sk_buff_head *xmitq)
 {
-	struct tipc_link **active = &n_ptr->active_links[0];
-	u32 i;
-	u32 highest_prio = 0;
+	tipc_node_lock(n);
+	__tipc_node_link_up(n, bearer_id, xmitq);
+	tipc_node_unlock(n);
+}
 
-	active[0] = active[1] = NULL;
+/**
+ * __tipc_node_link_down - handle loss of link
+ */
+static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id,
+				  struct sk_buff_head *xmitq,
+				  struct tipc_media_addr **maddr)
+{
+	struct tipc_link_entry *le = &n->links[*bearer_id];
+	int *slot0 = &n->active_links[0];
+	int *slot1 = &n->active_links[1];
+	int i, highest = 0;
+	struct tipc_link *l, *_l, *tnl;
+
+	l = n->links[*bearer_id].link;
+	if (!l || tipc_link_is_reset(l))
+		return;
 
-	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_DOWN;
+	n->link_id = l->peer_bearer_id << 16 | *bearer_id;
 
-		if (!l_ptr || !tipc_link_is_up(l_ptr) ||
-		    (l_ptr->priority < highest_prio))
-			continue;
+	tipc_bearer_remove_dest(n->net, *bearer_id, n->addr);
 
-		if (l_ptr->priority > highest_prio) {
-			highest_prio = l_ptr->priority;
-			active[0] = active[1] = l_ptr;
-		} else {
-			active[1] = l_ptr;
+	pr_debug("Lost link <%s> on network plane %c\n",
+		 l->name, l->net_plane);
+
+	/* 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 == l)
+			continue;
+		if (_l->priority < highest)
+			continue;
+		if (_l->priority > highest) {
+			highest = _l->priority;
+			*slot0 = i;
+			*slot1 = i;
+			continue;
 		}
+		*slot1 = i;
+	}
+
+	if (!tipc_node_is_up(n)) {
+		tipc_link_reset(l);
+		node_lost_contact(n, &le->inputq);
+		return;
 	}
+
+	/* There is still a working link => initiate failover */
+	tnl = node_active_link(n, 0);
+	tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT);
+	tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT);
+	n->sync_point = tnl->rcv_nxt + (U16_MAX / 2 - 1);
+	tipc_link_tnl_prepare(l, tnl, FAILOVER_MSG, xmitq);
+	tipc_link_reset(l);
+	tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT);
+	tipc_node_fsm_evt(n, NODE_FAILOVER_BEGIN_EVT);
+	*maddr = &n->links[tnl->bearer_id].maddr;
+	*bearer_id = tnl->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)
+static void tipc_node_link_down(struct tipc_node *n, int bearer_id, bool delete)
 {
-	struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id);
-	struct tipc_link **active;
+	struct tipc_link_entry *le = &n->links[bearer_id];
+	struct tipc_media_addr *maddr;
+	struct sk_buff_head xmitq;
+
+	__skb_queue_head_init(&xmitq);
+
+	tipc_node_lock(n);
+	__tipc_node_link_down(n, &bearer_id, &xmitq, &maddr);
+	if (delete && le->link) {
+		kfree(le->link);
+		le->link = NULL;
+		n->link_cnt--;
+	}
+	tipc_node_unlock(n);
 
-	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;
+	tipc_bearer_xmit(n->net, bearer_id, &xmitq, maddr);
+	tipc_sk_rcv(n->net, &le->inputq);
+}
 
-	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);
-
-	/* 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;
+bool tipc_node_is_up(struct tipc_node *n)
+{
+	return n->active_links[0] != INVALID_BEARER_ID;
+}
+
+void tipc_node_check_dest(struct net *net, u32 onode,
+			  struct tipc_bearer *b,
+			  u16 capabilities, u32 signature,
+			  struct tipc_media_addr *maddr,
+			  bool *respond, bool *dupl_addr)
+{
+	struct tipc_node *n;
+	struct tipc_link *l;
+	struct tipc_link_entry *le;
+	bool addr_match = false;
+	bool sign_match = false;
+	bool link_up = false;
+	bool accept_addr = false;
+	bool reset = true;
+
+	*dupl_addr = false;
+	*respond = false;
+
+	n = tipc_node_create(net, onode, capabilities);
+	if (!n)
 		return;
+
+	tipc_node_lock(n);
+
+	le = &n->links[b->identity];
+
+	/* Prepare to validate requesting node's signature and media address */
+	l = le->link;
+	link_up = l && tipc_link_is_up(l);
+	addr_match = l && !memcmp(&le->maddr, maddr, sizeof(*maddr));
+	sign_match = (signature == n->signature);
+
+	/* These three flags give us eight permutations: */
+
+	if (sign_match && addr_match && link_up) {
+		/* All is fine. Do nothing. */
+		reset = false;
+	} else if (sign_match && addr_match && !link_up) {
+		/* Respond. The link will come up in due time */
+		*respond = true;
+	} else if (sign_match && !addr_match && link_up) {
+		/* Peer has changed i/f address without rebooting.
+		 * If so, the link will reset soon, and the next
+		 * discovery will be accepted. So we can ignore it.
+		 * It may also be an cloned or malicious peer having
+		 * chosen the same node address and signature as an
+		 * existing one.
+		 * Ignore requests until the link goes down, if ever.
+		 */
+		*dupl_addr = true;
+	} else if (sign_match && !addr_match && !link_up) {
+		/* Peer link has changed i/f address without rebooting.
+		 * It may also be a cloned or malicious peer; we can't
+		 * distinguish between the two.
+		 * The signature is correct, so we must accept.
+		 */
+		accept_addr = true;
+		*respond = true;
+	} else if (!sign_match && addr_match && link_up) {
+		/* Peer node rebooted. Two possibilities:
+		 *  - Delayed re-discovery; this link endpoint has already
+		 *    reset and re-established contact with the peer, before
+		 *    receiving a discovery message from that node.
+		 *    (The peer happened to receive one from this node first).
+		 *  - The peer came back so fast that our side has not
+		 *    discovered it yet. Probing from this side will soon
+		 *    reset the link, since there can be no working link
+		 *    endpoint at the peer end, and the link will re-establish.
+		 *  Accept the signature, since it comes from a known peer.
+		 */
+		n->signature = signature;
+	} else if (!sign_match && addr_match && !link_up) {
+		/*  The peer node has rebooted.
+		 *  Accept signature, since it is a known peer.
+		 */
+		n->signature = signature;
+		*respond = true;
+	} else if (!sign_match && !addr_match && link_up) {
+		/* Peer rebooted with new address, or a new/duplicate peer.
+		 * Ignore until the link goes down, if ever.
+		 */
+		*dupl_addr = true;
+	} else if (!sign_match && !addr_match && !link_up) {
+		/* Peer rebooted with new address, or it is a new peer.
+		 * Accept signature and address.
+		 */
+		n->signature = signature;
+		accept_addr = true;
+		*respond = true;
 	}
-	/* 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;
+
+	if (!accept_addr)
+		goto exit;
+
+	/* Now create new link if not already existing */
+	if (!l) {
+		if (n->link_cnt == 2) {
+			pr_warn("Cannot establish 3rd link to %x\n", n->addr);
+			goto exit;
+		}
+		if (!tipc_link_create(n, b, mod(tipc_net(net)->random),
+				      tipc_own_addr(net), onode, &le->maddr,
+				      &le->inputq, &n->bclink.namedq, &l)) {
+			*respond = false;
+			goto exit;
+		}
+		tipc_link_reset(l);
+		if (n->state == NODE_FAILINGOVER)
+			tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT);
+		le->link = l;
+		n->link_cnt++;
+		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(&le->maddr, maddr, sizeof(*maddr));
+exit:
+	tipc_node_unlock(n);
+	if (reset)
+		tipc_node_link_down(n, b->identity, false);
+	tipc_node_put(n);
 }
 
-int tipc_node_active_links(struct tipc_node *n_ptr)
+void tipc_node_delete_links(struct net *net, int bearer_id)
 {
-	return n_ptr->active_links[0] != NULL;
+	struct tipc_net *tn = net_generic(net, tipc_net_id);
+	struct tipc_node *n;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(n, &tn->node_list, list) {
+		tipc_node_link_down(n, bearer_id, true);
+	}
+	rcu_read_unlock();
 }
 
-int tipc_node_is_up(struct tipc_node *n_ptr)
+static void tipc_node_reset_links(struct tipc_node *n)
 {
-	return tipc_node_active_links(n_ptr);
+	char addr_string[16];
+	int i;
+
+	pr_warn("Resetting all links to %s\n",
+		tipc_addr_string_fill(addr_string, n->addr));
+
+	for (i = 0; i < MAX_BEARERS; i++) {
+		tipc_node_link_down(n, i, false);
+	}
 }
 
-void tipc_node_attach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
+/* 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)
 {
-	n_ptr->links[l_ptr->bearer_id] = l_ptr;
-	n_ptr->link_cnt++;
+	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;
+		case NODE_SYNCH_END_EVT:
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_FAILOVER_BEGIN_EVT:
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_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 NODE_SYNCH_BEGIN_EVT:
+			state = NODE_SYNCHING;
+			break;
+		case NODE_FAILOVER_BEGIN_EVT:
+			state = NODE_FAILINGOVER;
+			break;
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+		case NODE_SYNCH_END_EVT:
+		case NODE_FAILOVER_END_EVT:
+			break;
+		default:
+			goto illegal_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;
+		case NODE_SYNCH_END_EVT:
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_FAILOVER_BEGIN_EVT:
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_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;
+		case NODE_SYNCH_END_EVT:
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_FAILOVER_BEGIN_EVT:
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_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;
+		case NODE_SYNCH_END_EVT:
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_FAILOVER_BEGIN_EVT:
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_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;
+		case NODE_SYNCH_END_EVT:
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_FAILOVER_BEGIN_EVT:
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_evt;
+		}
+		break;
+	case NODE_FAILINGOVER:
+		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 NODE_FAILOVER_END_EVT:
+			state = SELF_UP_PEER_UP;
+			break;
+		case NODE_FAILOVER_BEGIN_EVT:
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+			break;
+		case NODE_SYNCH_BEGIN_EVT:
+		case NODE_SYNCH_END_EVT:
+		default:
+			goto illegal_evt;
+		}
+		break;
+	case NODE_SYNCHING:
+		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 NODE_SYNCH_END_EVT:
+			state = SELF_UP_PEER_UP;
+			break;
+		case NODE_FAILOVER_BEGIN_EVT:
+			state = NODE_FAILINGOVER;
+			break;
+		case NODE_SYNCH_BEGIN_EVT:
+		case SELF_ESTABL_CONTACT_EVT:
+		case PEER_ESTABL_CONTACT_EVT:
+			break;
+		case NODE_FAILOVER_END_EVT:
+		default:
+			goto illegal_evt;
+		}
+		break;
+	default:
+		pr_err("Unknown node fsm state %x\n", state);
+		break;
+	}
+	n->state = state;
+	return;
+
+illegal_evt:
+	pr_err("Illegal node fsm evt %x in state %x\n", evt, state);
 }
 
-void tipc_node_detach_link(struct tipc_node *n_ptr, struct tipc_link *l_ptr)
+bool tipc_node_filter_pkt(struct tipc_node *n, struct tipc_msg *hdr)
 {
-	int i;
+	int state = n->state;
 
-	for (i = 0; i < MAX_BEARERS; i++) {
-		if (l_ptr != n_ptr->links[i])
-			continue;
-		n_ptr->links[i] = NULL;
-		n_ptr->link_cnt--;
+	if (likely(state == SELF_UP_PEER_UP))
+		return true;
+
+	if (state == SELF_LEAVING_PEER_DOWN)
+		return false;
+
+	if (state == SELF_DOWN_PEER_LEAVING) {
+		if (msg_peer_node_is_up(hdr))
+			return false;
 	}
+
+	return true;
 }
 
 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);
 	tipc_bclink_add_node(n_ptr->net, n_ptr->addr);
 }
 
-static void node_lost_contact(struct tipc_node *n_ptr)
+static void node_lost_contact(struct tipc_node *n_ptr,
+			      struct sk_buff_head *inputq)
 {
 	char addr_string[16];
 	struct tipc_sock_conn *conn, *safe;
+	struct tipc_link *l;
 	struct list_head *conns = &n_ptr->conn_sks;
 	struct sk_buff *skb;
 	struct tipc_net *tn = net_generic(n_ptr->net, tipc_net_id);
@@ -396,21 +844,13 @@ 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];
-		if (!l_ptr)
-			continue;
-		l_ptr->flags &= ~LINK_FAILINGOVER;
-		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);
+		l = n_ptr->links[i].link;
+		if (l)
+			tipc_link_fsm_evt(l, LINK_FAILOVER_END_EVT);
 	}
 
-	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;
@@ -421,10 +861,8 @@ static void node_lost_contact(struct tipc_node *n_ptr)
 				      SHORT_H_SIZE, 0, tn->own_addr,
 				      conn->peer_node, conn->port,
 				      conn->peer_port, TIPC_ERR_NO_NODE);
-		if (likely(skb)) {
-			skb_queue_tail(n_ptr->inputq, skb);
-			n_ptr->action_flags |= TIPC_MSG_EVT;
-		}
+		if (likely(skb))
+			skb_queue_tail(inputq, skb);
 		list_del(&conn->list);
 		kfree(conn);
 	}
@@ -453,7 +891,7 @@ 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;
@@ -471,27 +909,20 @@ void tipc_node_unlock(struct tipc_node *node)
 	u32 flags = node->action_flags;
 	u32 link_id = 0;
 	struct list_head *publ_list;
-	struct sk_buff_head *inputq = node->inputq;
-	struct sk_buff_head *namedq;
 
-	if (likely(!flags || (flags == TIPC_MSG_EVT))) {
-		node->action_flags = 0;
+	if (likely(!flags)) {
 		spin_unlock_bh(&node->lock);
-		if (flags == TIPC_MSG_EVT)
-			tipc_sk_rcv(net, inputq);
 		return;
 	}
 
 	addr = node->addr;
 	link_id = node->link_id;
-	namedq = node->namedq;
 	publ_list = &node->publ_list;
 
-	node->action_flags &= ~(TIPC_MSG_EVT |
-				TIPC_NOTIFY_NODE_DOWN | TIPC_NOTIFY_NODE_UP |
+	node->action_flags &= ~(TIPC_NOTIFY_NODE_DOWN | TIPC_NOTIFY_NODE_UP |
 				TIPC_NOTIFY_LINK_DOWN | TIPC_NOTIFY_LINK_UP |
 				TIPC_WAKEUP_BCAST_USERS | TIPC_BCAST_MSG_EVT |
-				TIPC_NAMED_MSG_EVT | TIPC_BCAST_RESET);
+				TIPC_BCAST_RESET);
 
 	spin_unlock_bh(&node->lock);
 
@@ -512,17 +943,11 @@ void tipc_node_unlock(struct tipc_node *node)
 		tipc_nametbl_withdraw(net, TIPC_LINK_STATE, addr,
 				      link_id, addr);
 
-	if (flags & TIPC_MSG_EVT)
-		tipc_sk_rcv(net, inputq);
-
-	if (flags & TIPC_NAMED_MSG_EVT)
-		tipc_named_rcv(net, namedq);
-
 	if (flags & TIPC_BCAST_MSG_EVT)
 		tipc_bclink_input(net);
 
 	if (flags & TIPC_BCAST_RESET)
-		tipc_link_reset_all(node);
+		tipc_node_reset_links(node);
 }
 
 /* Caller should hold node lock for the passed node */
@@ -559,6 +984,290 @@ msg_full:
 	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);
+		tipc_node_unlock(n);
+		if (unlikely(rc == -ENOBUFS))
+			tipc_node_link_down(n, bearer_id, false);
+		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_node_check_state - check and if necessary update node state
+ * @skb: TIPC packet
+ * @bearer_id: identity of bearer delivering the packet
+ * Returns true if state is ok, otherwise consumes buffer and returns false
+ */
+static bool tipc_node_check_state(struct tipc_node *n, struct sk_buff *skb,
+				  int bearer_id, struct sk_buff_head *xmitq)
+{
+	struct tipc_msg *hdr = buf_msg(skb);
+	int usr = msg_user(hdr);
+	int mtyp = msg_type(hdr);
+	u16 oseqno = msg_seqno(hdr);
+	u16 iseqno = msg_seqno(msg_get_wrapped(hdr));
+	u16 exp_pkts = msg_msgcnt(hdr);
+	u16 rcv_nxt, syncpt, dlv_nxt;
+	int state = n->state;
+	struct tipc_link *l, *tnl, *pl = NULL;
+	struct tipc_media_addr *maddr;
+	int i, pb_id;
+
+	l = n->links[bearer_id].link;
+	if (!l)
+		return false;
+	rcv_nxt = l->rcv_nxt;
+
+
+	if (likely((state == SELF_UP_PEER_UP) && (usr != TUNNEL_PROTOCOL)))
+		return true;
+
+	/* Find parallel link, if any */
+	for (i = 0; i < MAX_BEARERS; i++) {
+		if ((i != bearer_id) && n->links[i].link) {
+			pl = n->links[i].link;
+			break;
+		}
+	}
+
+	/* Update node accesibility if applicable */
+	if (state == SELF_UP_PEER_COMING) {
+		if (!tipc_link_is_up(l))
+			return true;
+		if (!msg_peer_link_is_up(hdr))
+			return true;
+		tipc_node_fsm_evt(n, PEER_ESTABL_CONTACT_EVT);
+	}
+
+	if (state == SELF_DOWN_PEER_LEAVING) {
+		if (msg_peer_node_is_up(hdr))
+			return false;
+		tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT);
+	}
+
+	/* Ignore duplicate packets */
+	if (less(oseqno, rcv_nxt))
+		return true;
+
+	/* Initiate or update failover mode if applicable */
+	if ((usr == TUNNEL_PROTOCOL) && (mtyp == FAILOVER_MSG)) {
+		syncpt = oseqno + exp_pkts - 1;
+		if (pl && tipc_link_is_up(pl)) {
+			pb_id = pl->bearer_id;
+			__tipc_node_link_down(n, &pb_id, xmitq, &maddr);
+			tipc_skb_queue_splice_tail_init(pl->inputq, l->inputq);
+		}
+		/* If pkts arrive out of order, use lowest calculated syncpt */
+		if (less(syncpt, n->sync_point))
+			n->sync_point = syncpt;
+	}
+
+	/* Open parallel link when tunnel link reaches synch point */
+	if ((n->state == NODE_FAILINGOVER) && tipc_link_is_up(l)) {
+		if (!more(rcv_nxt, n->sync_point))
+			return true;
+		tipc_node_fsm_evt(n, NODE_FAILOVER_END_EVT);
+		if (pl)
+			tipc_link_fsm_evt(pl, LINK_FAILOVER_END_EVT);
+		return true;
+	}
+
+	/* No synching needed if only one link */
+	if (!pl || !tipc_link_is_up(pl))
+		return true;
+
+	/* Initiate or update synch mode if applicable */
+	if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG)) {
+		syncpt = iseqno + exp_pkts - 1;
+		if (!tipc_link_is_up(l)) {
+			tipc_link_fsm_evt(l, LINK_ESTABLISH_EVT);
+			__tipc_node_link_up(n, bearer_id, xmitq);
+		}
+		if (n->state == SELF_UP_PEER_UP) {
+			n->sync_point = syncpt;
+			tipc_link_fsm_evt(l, LINK_SYNCH_BEGIN_EVT);
+			tipc_node_fsm_evt(n, NODE_SYNCH_BEGIN_EVT);
+		}
+		if (less(syncpt, n->sync_point))
+			n->sync_point = syncpt;
+	}
+
+	/* Open tunnel link when parallel link reaches synch point */
+	if ((n->state == NODE_SYNCHING) && tipc_link_is_synching(l)) {
+		if (tipc_link_is_synching(l)) {
+			tnl = l;
+		} else {
+			tnl = pl;
+			pl = l;
+		}
+		dlv_nxt = pl->rcv_nxt - mod(skb_queue_len(pl->inputq));
+		if (more(dlv_nxt, n->sync_point)) {
+			tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT);
+			tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT);
+			return true;
+		}
+		if (l == pl)
+			return true;
+		if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG))
+			return true;
+		if (usr == LINK_PROTOCOL)
+			return true;
+		return false;
+	}
+	return true;
+}
+
+/**
+ * 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_msg *hdr = buf_msg(skb);
+	int usr = msg_user(hdr);
+	int bearer_id = b->identity;
+	struct tipc_link_entry *le;
+	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 */
+	if (unlikely(msg_non_seq(hdr))) {
+		if (usr ==  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;
+	le = &n->links[bearer_id];
+
+	tipc_node_lock(n);
+
+	/* Is reception permitted at the moment ? */
+	if (!tipc_node_filter_pkt(n, hdr))
+		goto unlock;
+
+	if (unlikely(msg_user(hdr) == LINK_PROTOCOL))
+		tipc_bclink_sync_state(n, hdr);
+
+	/* Release acked broadcast packets */
+	if (unlikely(n->bclink.acked != msg_bcast_ack(hdr)))
+		tipc_bclink_acknowledge(n, msg_bcast_ack(hdr));
+
+	/* Check and if necessary update node state */
+	if (likely(tipc_node_check_state(n, skb, bearer_id, &xmitq))) {
+		rc = tipc_link_rcv(le->link, skb, &xmitq);
+		skb = NULL;
+	}
+unlock:
+	tipc_node_unlock(n);
+
+	if (unlikely(rc & TIPC_LINK_UP_EVT))
+		tipc_node_link_up(n, bearer_id, &xmitq);
+
+	if (unlikely(rc & TIPC_LINK_DOWN_EVT))
+		tipc_node_link_down(n, bearer_id, false);
+
+	if (unlikely(!skb_queue_empty(&n->bclink.namedq)))
+		tipc_named_rcv(net, &n->bclink.namedq);
+
+	if (!skb_queue_empty(&le->inputq))
+		tipc_sk_rcv(net, &le->inputq);
+
+	if (!skb_queue_empty(&xmitq))
+		tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr);
+
+	tipc_node_put(n);
+discard:
+	kfree_skb(skb);
+}
+
 int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	int err;