tipc: add option to suppress PUBLISH events for pre-existing publications

Currently, when a user is subscribing for binding table publications,
he will receive a PUBLISH event for all already existing matching items
in the binding table.

However, a group socket making a subscriptions doesn't need this initial
status update from the binding table, because it has already scanned it
during the join operation. Worse, the multiplicatory effect of issuing
mutual events for dozens or hundreds group members within a short time
frame put a heavy load on the topology server, with the end result that
scale out operations on a big group tend to take much longer than needed.

We now add a new filter option, TIPC_SUB_NO_STATUS, for topology server
subscriptions, so that this initial avalanche of events is suppressed.
This change, along with the previous commit, significantly improves the
range and speed of group scale out operations.

We keep the new option internal for the tipc driver, at least for now.

Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 21 insertions, 15 deletions

diff --git a/net/tipc/group.c b/net/tipc/group.c
index 6ca07f0da60c..cf996bd6ec98 100644
--- a/net/tipc/group.c
+++ b/net/tipc/group.c
@@ -177,7 +177,9 @@ struct tipc_group *tipc_group_create(struct net *net, u32 portid,
         grp->scope = mreq->scope;
         grp->loopback = mreq->flags & TIPC_GROUP_LOOPBACK;
         grp->events = mreq->flags & TIPC_GROUP_MEMBER_EVTS;
-        if (tipc_topsrv_kern_subscr(net, portid, type, 0, ~0, &grp->subid))
+        if (tipc_topsrv_kern_subscr(net, portid, type,
+                                    TIPC_SUB_PORTS | TIPC_SUB_NO_STATUS,
+                                    0, ~0, &grp->subid))
                 return grp;
         kfree(grp);
         return NULL;
diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c
index e04ab72f313c..60af9885f160 100644
--- a/net/tipc/name_table.c
+++ b/net/tipc/name_table.c
@@ -405,12 +405,13 @@ found:
 }
 
 /**
- * tipc_nameseq_subscribe - attach a subscription, and issue
- * the prescribed number of events if there is any sub-
+ * tipc_nameseq_subscribe - attach a subscription, and optionally
+ * issue the prescribed number of events if there is any sub-
  * sequence overlapping with the requested sequence
  */
 static void tipc_nameseq_subscribe(struct name_seq *nseq,
-                                   struct tipc_subscription *s)
+                                   struct tipc_subscription *s,
+                                   bool status)
 {
         struct sub_seq *sseq = nseq->sseqs;
         struct tipc_name_seq ns;
@@ -420,7 +421,7 @@ static void tipc_nameseq_subscribe(struct name_seq *nseq,
         tipc_subscrp_get(s);
         list_add(&s->nameseq_list, &nseq->subscriptions);
 
-        if (!sseq)
+        if (!status || !sseq)
                 return;
 
         while (sseq != &nseq->sseqs[nseq->first_free]) {
@@ -811,7 +812,7 @@ int tipc_nametbl_withdraw(struct net *net, u32 type, u32 lower, u32 ref,
 /**
  * tipc_nametbl_subscribe - add a subscription object to the name table
  */
-void tipc_nametbl_subscribe(struct tipc_subscription *s)
+void tipc_nametbl_subscribe(struct tipc_subscription *s, bool status)
 {
         struct tipc_net *tn = net_generic(s->net, tipc_net_id);
         u32 type = tipc_subscrp_convert_seq_type(s->evt.s.seq.type, s->swap);
@@ -825,7 +826,7 @@ void tipc_nametbl_subscribe(struct tipc_subscription *s)
                 seq = tipc_nameseq_create(type, &tn->nametbl->seq_hlist[index]);
         if (seq) {
                 spin_lock_bh(&seq->lock);
-                tipc_nameseq_subscribe(seq, s);
+                tipc_nameseq_subscribe(seq, s, status);
                 spin_unlock_bh(&seq->lock);
         } else {
                 tipc_subscrp_convert_seq(&s->evt.s.seq, s->swap, &ns);
diff --git a/net/tipc/name_table.h b/net/tipc/name_table.h
index 71926e429446..73a148c85c15 100644
--- a/net/tipc/name_table.h
+++ b/net/tipc/name_table.h
@@ -121,7 +121,7 @@ struct publication *tipc_nametbl_insert_publ(struct net *net, u32 type,
 struct publication *tipc_nametbl_remove_publ(struct net *net, u32 type,
                                              u32 lower, u32 node, u32 ref,
                                              u32 key);
-void tipc_nametbl_subscribe(struct tipc_subscription *s);
+void tipc_nametbl_subscribe(struct tipc_subscription *s, bool status);
 void tipc_nametbl_unsubscribe(struct tipc_subscription *s);
 int tipc_nametbl_init(struct net *net);
 void tipc_nametbl_stop(struct net *net);
diff --git a/net/tipc/server.c b/net/tipc/server.c
index d60c30342327..950c54cbcf3a 100644
--- a/net/tipc/server.c
+++ b/net/tipc/server.c
@@ -490,7 +490,7 @@ void tipc_conn_terminate(struct tipc_server *s, int conid)
 }
 
 bool tipc_topsrv_kern_subscr(struct net *net, u32 port, u32 type,
-                             u32 lower, u32 upper, int *conid)
+                             u32 filter, u32 lower, u32 upper, int *conid)
 {
         struct tipc_subscriber *scbr;
         struct tipc_subscr sub;
@@ -501,7 +501,7 @@ bool tipc_topsrv_kern_subscr(struct net *net, u32 port, u32 type,
         sub.seq.lower = lower;
         sub.seq.upper = upper;
         sub.timeout = TIPC_WAIT_FOREVER;
-        sub.filter = TIPC_SUB_PORTS;
+        sub.filter = filter;
         *(u32 *)&sub.usr_handle = port;
 
         con = tipc_alloc_conn(tipc_topsrv(net));
diff --git a/net/tipc/server.h b/net/tipc/server.h
index 2113c9192633..ea1effbff23e 100644
--- a/net/tipc/server.h
+++ b/net/tipc/server.h
@@ -41,6 +41,7 @@
 #include <net/net_namespace.h>
 
 #define TIPC_SERVER_NAME_LEN    32
+#define TIPC_SUB_NO_STATUS      0x80
 
 /**
  * struct tipc_server - TIPC server structure
@@ -84,7 +85,7 @@ int tipc_conn_sendmsg(struct tipc_server *s, int conid,
                       struct sockaddr_tipc *addr, void *data, size_t len);
 
 bool tipc_topsrv_kern_subscr(struct net *net, u32 port, u32 type,
-                             u32 lower, u32 upper, int *conid);
+                             u32 filter, u32 lower, u32 upper, int *conid);
 void tipc_topsrv_kern_unsubscr(struct net *net, int conid);
 
 /**
diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c
index 251065dfd8df..1052341a0ea9 100644
--- a/net/tipc/subscr.c
+++ b/net/tipc/subscr.c
@@ -286,7 +286,8 @@ static struct tipc_subscription *tipc_subscrp_create(struct net *net,
 }
 
 static void tipc_subscrp_subscribe(struct net *net, struct tipc_subscr *s,
-                                   struct tipc_subscriber *subscriber, int swap)
+                                   struct tipc_subscriber *subscriber, int swap,
+                                   bool status)
 {
         struct tipc_net *tn = net_generic(net, tipc_net_id);
         struct tipc_subscription *sub = NULL;
@@ -299,7 +300,7 @@ static void tipc_subscrp_subscribe(struct net *net, struct tipc_subscr *s,
         spin_lock_bh(&subscriber->lock);
         list_add(&sub->subscrp_list, &subscriber->subscrp_list);
         sub->subscriber = subscriber;
-        tipc_nametbl_subscribe(sub);
+        tipc_nametbl_subscribe(sub, status);
         tipc_subscrb_get(subscriber);
         spin_unlock_bh(&subscriber->lock);
 
@@ -323,6 +324,7 @@ static void tipc_subscrb_rcv_cb(struct net *net, int conid,
 {
         struct tipc_subscriber *subscriber = usr_data;
         struct tipc_subscr *s = (struct tipc_subscr *)buf;
+        bool status;
         int swap;
 
         /* Determine subscriber's endianness */
@@ -334,8 +336,8 @@ static void tipc_subscrb_rcv_cb(struct net *net, int conid,
                 s->filter &= ~htohl(TIPC_SUB_CANCEL, swap);
                 return tipc_subscrp_cancel(s, subscriber);
         }
-
-        tipc_subscrp_subscribe(net, s, subscriber, swap);
+        status = !(s->filter & htohl(TIPC_SUB_NO_STATUS, swap));
+        tipc_subscrp_subscribe(net, s, subscriber, swap, status);
 }
 
 /* Handle one request to establish a new subscriber */