tipc: fix failover problem

We see the following scenario:
1) Link endpoint B on node 1 discovers that its peer endpoint is gone.
   Since there is a second working link, failover procedure is started.
2) Link endpoint A on node 1 sends a FAILOVER message to peer endpoint
   A on node 2. The node item 1->2 goes to state FAILINGOVER.
3) Linke endpoint A/2 receives the failover, and is supposed to take
   down its parallell link endpoint B/2, while producing a FAILOVER
   message to send back to A/1.
4) However, B/2 has already been deleted, so no FAILOVER message can
   created.
5) Node 1->2 remains in state FAILINGOVER forever, refusing to receive
   any messages that can bring B/1 up again. We are left with a non-
   redundant link between node 1 and 2.

We fix this with letting endpoint A/2 build a dummy FAILOVER message
to send to back to A/1, so that the situation can be resolved.

Signed-off-by: LUU Duc Canh <canh.d.luu@dektech.com.au>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/net/tipc/link.c b/net/tipc/link.c
index 26cc033ee167591d186ecfbb5a6fbbb87d6c6d10..4ed650ce6e618e3bc10849cead6750368bfa47eb 100644 (file)
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -410,6 +410,11 @@ char *tipc_link_name(struct tipc_link *l)
        return l->name;
 }
 
+u32 tipc_link_state(struct tipc_link *l)
+{
+       return l->state;
+}
+
 /**
  * tipc_link_create - create a new link
  * @n: pointer to associated node
@@ -1385,6 +1390,36 @@ static void tipc_link_build_proto_msg(struct tipc_link *l, int mtyp, bool probe,
        __skb_queue_tail(xmitq, skb);
 }
 
+void tipc_link_create_dummy_tnl_msg(struct tipc_link *l,
+                                   struct sk_buff_head *xmitq)
+{
+       u32 onode = tipc_own_addr(l->net);
+       struct tipc_msg *hdr, *ihdr;
+       struct sk_buff_head tnlq;
+       struct sk_buff *skb;
+       u32 dnode = l->addr;
+
+       skb_queue_head_init(&tnlq);
+       skb = tipc_msg_create(TUNNEL_PROTOCOL, FAILOVER_MSG,
+                             INT_H_SIZE, BASIC_H_SIZE,
+                             dnode, onode, 0, 0, 0);
+       if (!skb) {
+               pr_warn("%sunable to create tunnel packet\n", link_co_err);
+               return;
+       }
+
+       hdr = buf_msg(skb);
+       msg_set_msgcnt(hdr, 1);
+       msg_set_bearer_id(hdr, l->peer_bearer_id);
+
+       ihdr = (struct tipc_msg *)msg_data(hdr);
+       tipc_msg_init(onode, ihdr, TIPC_LOW_IMPORTANCE, TIPC_DIRECT_MSG,
+                     BASIC_H_SIZE, dnode);
+       msg_set_errcode(ihdr, TIPC_ERR_NO_PORT);
+       __skb_queue_tail(&tnlq, skb);
+       tipc_link_xmit(l, &tnlq, xmitq);
+}
+
 /* tipc_link_tnl_prepare(): prepare and return a list of tunnel packets
  * with contents of the link's transmit and backlog queues.
  */

diff --git a/net/tipc/link.h b/net/tipc/link.h
index 7bc494a33fdf1c3cdf8feb04b44db7e6e04a349c..90488c538a4e4edaddfaf441a517f8b29c1d2ebc 100644 (file)
--- a/net/tipc/link.h
+++ b/net/tipc/link.h
@@ -88,6 +88,8 @@ bool tipc_link_bc_create(struct net *net, u32 ownnode, u32 peer,
                         struct tipc_link **link);
 void tipc_link_tnl_prepare(struct tipc_link *l, struct tipc_link *tnl,
                           int mtyp, struct sk_buff_head *xmitq);
+void tipc_link_create_dummy_tnl_msg(struct tipc_link *tnl,
+                                   struct sk_buff_head *xmitq);
 void tipc_link_build_reset_msg(struct tipc_link *l, struct sk_buff_head *xmitq);
 int tipc_link_fsm_evt(struct tipc_link *l, int evt);
 bool tipc_link_is_up(struct tipc_link *l);
@@ -107,6 +109,7 @@ u16 tipc_link_rcv_nxt(struct tipc_link *l);
 u16 tipc_link_acked(struct tipc_link *l);
 u32 tipc_link_id(struct tipc_link *l);
 char *tipc_link_name(struct tipc_link *l);
+u32 tipc_link_state(struct tipc_link *l);
 char tipc_link_plane(struct tipc_link *l);
 int tipc_link_prio(struct tipc_link *l);
 int tipc_link_window(struct tipc_link *l);

diff --git a/net/tipc/node.c b/net/tipc/node.c
index 68014f1b69765269236ac0a6d839ca754d1235da..b0ee25f1f2e6216bb05c18373aee455872ff5f14 100644 (file)
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -111,6 +111,7 @@ struct tipc_node {
        int action_flags;
        struct list_head list;
        int state;
+       bool failover_sent;
        u16 sync_point;
        int link_cnt;
        u16 working_links;
@@ -680,6 +681,7 @@ static void __tipc_node_link_up(struct tipc_node *n, int bearer_id,
                *slot0 = bearer_id;
                *slot1 = bearer_id;
                tipc_node_fsm_evt(n, SELF_ESTABL_CONTACT_EVT);
+               n->failover_sent = false;
                n->action_flags |= TIPC_NOTIFY_NODE_UP;
                tipc_link_set_active(nl, true);
                tipc_bcast_add_peer(n->net, nl, xmitq);
@@ -1615,6 +1617,15 @@ static bool tipc_node_check_state(struct tipc_node *n, struct sk_buff *skb,
                        tipc_skb_queue_splice_tail_init(tipc_link_inputq(pl),
                                                        tipc_link_inputq(l));
                }
+               /* If parallel link was already down, and this happened before
+                * the tunnel link came up, FAILOVER was never sent. Ensure that
+                * FAILOVER is sent to get peer out of NODE_FAILINGOVER state.
+                */
+               if (n->state != NODE_FAILINGOVER && !n->failover_sent) {
+                       tipc_link_create_dummy_tnl_msg(l, xmitq);
+                       n->failover_sent = true;
+               }
+
                /* If pkts arrive out of order, use lowest calculated syncpt */
                if (less(syncpt, n->sync_point))
                        n->sync_point = syncpt;