rdma: glusterd crash if rdma_disconnect is called as soon as connect a request.

        Backport of http://review.gluster.org/8925

we are initializing connection in server side immediately after
rdma_accept is called. But we are delaying adding the transport
to listener list until getting RDMA_CM_EVENT_ESTABLISHED event.
Before getting this event if disconnect is called glusterd will
try to remove the transport from list which is not added. So if
the list is empty it causes a glusterd crash . In this patch we
will call the function to initialize the connection as soon as
rdma_accept is called.

Change-Id: Ie783fdbb4342459e5bc162bf8600bf47c1e2e909
BUG: 1166515
Signed-off-by: Mohammed Rafi KC <rkavunga@redhat.com>
Reviewed-on: http://review.gluster.org/8925
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Tested-by: Raghavendra G <rgowdapp@redhat.com>
Reviewed-on: http://review.gluster.org/9174
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra Bhat <raghavendra@redhat.com>

1 files changed, 17 insertions, 9 deletions

diff --git a/rpc/rpc-transport/rdma/src/rdma.c b/rpc/rpc-transport/rdma/src/rdma.c
index 4837a17532e..e65713a4893 100644
--- a/rpc/rpc-transport/rdma/src/rdma.c
+++ b/rpc/rpc-transport/rdma/src/rdma.c
@@ -58,6 +58,8 @@ gf_rdma_disconnect (rpc_transport_t *this);
 static void
 gf_rdma_cm_handle_disconnect (rpc_transport_t *this);
 
+static int
+gf_rdma_cm_handle_connect_init (struct rdma_cm_event *event);
 
 static void
 gf_rdma_put_post (gf_rdma_queue_t *queue, gf_rdma_post_t *post)
@@ -785,7 +787,7 @@ gf_rdma_cm_handle_connect_request (struct rdma_cm_event *event)
         rpc_transport_t        *this        = NULL, *listener = NULL;
         struct rdma_cm_id      *child_cm_id = NULL, *listener_cm_id = NULL;
         struct rdma_conn_param  conn_param  = {0, };
-        gf_rdma_private_t      *priv        = NULL, *child_priv = NULL;
+        gf_rdma_private_t      *priv        = NULL;
         gf_rdma_options_t      *options     = NULL;
 
         child_cm_id = event->id;
@@ -804,7 +806,7 @@ gf_rdma_cm_handle_connect_request (struct rdma_cm_event *event)
                 rdma_destroy_id (child_cm_id);
                 goto out;
         }
-        child_priv = this->private;
+
         gf_log (listener->name, GF_LOG_TRACE,
                 "got a connect request (me:%s peer:%s)",
                 listener->myinfo.identifier, this->peerinfo.identifier);
@@ -831,7 +833,7 @@ gf_rdma_cm_handle_connect_request (struct rdma_cm_event *event)
                 gf_rdma_cm_handle_disconnect (this);
                 goto out;
         }
-        child_priv->connected = 1;
+        gf_rdma_cm_handle_connect_init (event);
         ret = 0;
 
 out:
@@ -965,7 +967,7 @@ gf_rdma_cm_handle_disconnect (rpc_transport_t *this)
 
 
 static int
-gf_rdma_cm_handle_event_established (struct rdma_cm_event *event)
+gf_rdma_cm_handle_connect_init (struct rdma_cm_event *event)
 {
         rpc_transport_t   *this  = NULL;
         gf_rdma_private_t *priv  = NULL;
@@ -976,6 +978,13 @@ gf_rdma_cm_handle_event_established (struct rdma_cm_event *event)
         this = cm_id->context;
         priv = this->private;
 
+        if (priv->connected == 1) {
+                gf_log (this->name, GF_LOG_TRACE,
+                        "received event RDMA_CM_EVENT_ESTABLISHED (me:%s peer:%s)",
+                        this->myinfo.identifier, this->peerinfo.identifier);
+                return ret;
+        }
+
         priv->connected = 1;
 
         pthread_mutex_lock (&priv->write_mutex);
@@ -986,6 +995,9 @@ gf_rdma_cm_handle_event_established (struct rdma_cm_event *event)
         pthread_mutex_unlock (&priv->write_mutex);
 
         if (priv->entity == GF_RDMA_CLIENT) {
+                gf_log (this->name, GF_LOG_TRACE,
+                        "received event RDMA_CM_EVENT_ESTABLISHED (me:%s peer:%s)",
+                        this->myinfo.identifier, this->peerinfo.identifier);
                 ret = rpc_transport_notify (this, RPC_TRANSPORT_CONNECT, this);
 
         } else if (priv->entity == GF_RDMA_SERVER) {
@@ -997,10 +1009,6 @@ gf_rdma_cm_handle_event_established (struct rdma_cm_event *event)
                 gf_rdma_disconnect (this);
         }
 
-        gf_log (this->name, GF_LOG_TRACE,
-                "received event RDMA_CM_EVENT_ESTABLISHED (me:%s peer:%s)",
-                this->myinfo.identifier, this->peerinfo.identifier);
-
         return ret;
 }
 
@@ -1060,7 +1068,7 @@ gf_rdma_cm_event_handler (void *data)
                         break;
 
                 case RDMA_CM_EVENT_ESTABLISHED:
-                        gf_rdma_cm_handle_event_established (event);
+                        gf_rdma_cm_handle_connect_init (event);
                         break;
 
                 case RDMA_CM_EVENT_ADDR_ERROR: