From: Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com>
Date: Fri, 18 Sep 2015 01:37:01 +0000 (-0700)
Subject: RDS: Handle RDMA_CM_EVENT_TIMEWAIT_EXIT event.
X-Git-Tag: v4.1.12-92~264^2~7^2
X-Git-Url: https://www.infradead.org/git/?a=commitdiff_plain;h=d109c1c0637711d6a42ca90ba1fb2bce18b4fe71;p=users%2Fjedix%2Flinux-maple.git

RDS: Handle RDMA_CM_EVENT_TIMEWAIT_EXIT event.

RDS/IB connection can hang with the log message
   "CQ access violation on CQN ..."
followed by
   "RDS: unknown event 15!".

Event 15 is RDMA_CM_EVENT_TIMEWAIT_EXIT. RDS was not handling this.
With this fix RDS will now attempt to reconnect on getting this event.

The fix contains 2 changes.
1) RDS change to handle RDMA_CM_EVENT_TIMEWAIT_EXIT event.
2) Display diagnostic data of "syndrome" and "vendor_error_syndrome" in
   mlx4_core when CQ access violation occurs.

Orabug: 21675221

Acked-by: Chien Yen <chien.yen@oracle.com>
Signed-off-by: Venkat Venkatsubra <venkat.x.venkatsubra@oracle.com>
Signed-off-by: Mukesh Kacker <mukesh.kacker@oracle.com>
---

diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c
index 2619c9fbf42df..5677d8a1c51ea 100644
--- a/drivers/net/ethernet/mellanox/mlx4/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -627,10 +627,13 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)
 		}
 
 		case MLX4_EVENT_TYPE_CQ_ERROR:
-			mlx4_warn(dev, "CQ %s on CQN %06x\n",
+			mlx4_warn(dev,
+		 "CQ %s on CQN %06x syndrome=0x%x vendor_error_syndrome=0x%x\n",
 				  eqe->event.cq_err.syndrome == 1 ?
 				  "overrun" : "access violation",
-				  be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff);
+				  be32_to_cpu(eqe->event.cq_err.cqn) & 0xffffff,
+				  eqe->event.cq_err.syndrome,
+				  eqe->event.cq_err.reserved2[2]);
 			if (mlx4_is_master(dev)) {
 				ret = mlx4_get_slave_from_resource_id(dev,
 					RES_CQ,
diff --git a/net/rds/rdma_transport.c b/net/rds/rdma_transport.c
index b826fcc9c709e..42dad98f6cbe4 100644
--- a/net/rds/rdma_transport.c
+++ b/net/rds/rdma_transport.c
@@ -276,6 +276,17 @@ int rds_rdma_cm_event_handler(struct rdma_cm_id *cm_id,
 		rds_conn_drop(conn);
 		break;
 
+	case RDMA_CM_EVENT_TIMEWAIT_EXIT:
+		if (conn) {
+			printk(KERN_INFO "TIMEWAIT_EXIT event - "
+				"dropping connection "
+				"%pI4->%pI4\n", &conn->c_laddr,
+				 &conn->c_faddr);
+			rds_conn_drop(conn);
+		} else
+			printk(KERN_INFO "TIMEWAIT_EXIT event - conn=NULL\n");
+		break;
+
 	default:
 		/* things like device disconnect? */
 		printk(KERN_ERR "RDS: unknown event %u!\n", event->event);