RDS: Base init_depth and responder_resources on hw values

Instead of using a constant for initiator_depth and
responder_resources, read the per-QP values when the
device is enumerated, and then use these values when creating
the connection.

Signed-off-by: Andy Grover <andy.grover@oracle.com>

4 files changed, 24 insertions, 11 deletions

diff --git a/net/rds/ib.c b/net/rds/ib.c
index f0d2965..72a5116 100644
--- a/net/rds/ib.c
+++ b/net/rds/ib.c
@@ -91,6 +91,9 @@ void rds_ib_add_one(struct ib_device *device)
 			min_t(unsigned int, dev_attr->max_fmr, fmr_pool_size) :
 			fmr_pool_size;
 
+	rds_ibdev->max_initiator_depth = dev_attr->max_qp_init_rd_atom;
+	rds_ibdev->max_responder_resources = dev_attr->max_qp_rd_atom;
+
 	rds_ibdev->dev = device;
 	rds_ibdev->pd = ib_alloc_pd(device);
 	if (IS_ERR(rds_ibdev->pd))
diff --git a/net/rds/ib.h b/net/rds/ib.h
index d2fd0aa..a303f13 100644
--- a/net/rds/ib.h
+++ b/net/rds/ib.h
@@ -164,6 +164,8 @@ struct rds_ib_device {
 	unsigned int		max_fmrs;
 	int			max_sge;
 	unsigned int		max_wrs;
+	unsigned int		max_initiator_depth;
+	unsigned int		max_responder_resources;
 	spinlock_t		spinlock;	/* protect the above */
 };
 
diff --git a/net/rds/ib_cm.c b/net/rds/ib_cm.c
index b46bc2f..3134336 100644
--- a/net/rds/ib_cm.c
+++ b/net/rds/ib_cm.c
@@ -153,18 +153,25 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
 static void rds_ib_cm_fill_conn_param(struct rds_connection *conn,
 			struct rdma_conn_param *conn_param,
 			struct rds_ib_connect_private *dp,
-			u32 protocol_version)
+			u32 protocol_version,
+			u32 max_responder_resources,
+			u32 max_initiator_depth)
 {
+	struct rds_ib_connection *ic = conn->c_transport_data;
+	struct rds_ib_device *rds_ibdev;
+
 	memset(conn_param, 0, sizeof(struct rdma_conn_param));
-	/* XXX tune these? */
-	conn_param->responder_resources = 1;
-	conn_param->initiator_depth = 1;
+
+	rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
+
+	conn_param->responder_resources =
+		min_t(u32, rds_ibdev->max_responder_resources, max_responder_resources);
+	conn_param->initiator_depth =
+		min_t(u32, rds_ibdev->max_initiator_depth, max_initiator_depth);
 	conn_param->retry_count = min_t(unsigned int, rds_ib_retry_count, 7);
 	conn_param->rnr_retry_count = 7;
 
 	if (dp) {
-		struct rds_ib_connection *ic = conn->c_transport_data;
-
 		memset(dp, 0, sizeof(*dp));
 		dp->dp_saddr = conn->c_laddr;
 		dp->dp_daddr = conn->c_faddr;
@@ -479,7 +486,9 @@ int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
 		goto out;
 	}
 
-	rds_ib_cm_fill_conn_param(conn, &conn_param, &dp_rep, version);
+	rds_ib_cm_fill_conn_param(conn, &conn_param, &dp_rep, version,
+		event->param.conn.responder_resources,
+		event->param.conn.initiator_depth);
 
 	/* rdma_accept() calls rdma_reject() internally if it fails */
 	err = rdma_accept(cm_id, &conn_param);
@@ -516,8 +525,8 @@ int rds_ib_cm_initiate_connect(struct rdma_cm_id *cm_id)
 		goto out;
 	}
 
-	rds_ib_cm_fill_conn_param(conn, &conn_param, &dp, RDS_PROTOCOL_VERSION);
-
+	rds_ib_cm_fill_conn_param(conn, &conn_param, &dp, RDS_PROTOCOL_VERSION,
+		UINT_MAX, UINT_MAX);
 	ret = rdma_connect(cm_id, &conn_param);
 	if (ret)
 		rds_ib_conn_error(conn, "rdma_connect failed (%d)\n", ret);
diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index a7019df..abbc297 100644
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -745,7 +745,6 @@ int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
 		rm->atomic.op_swap_add = args->fadd.add;
 	}
 
-	rm->m_rdma_cookie = args->cookie;
 	rm->atomic.op_notify = !!(args->flags & RDS_RDMA_NOTIFY_ME);
 	rm->atomic.op_recverr = rs->rs_recverr;
 	rm->atomic.op_sg = rds_message_alloc_sgs(rm, 1);
@@ -779,7 +778,7 @@ int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
 		rm->atomic.op_notifier->n_status = RDS_RDMA_SUCCESS;
 	}
 
-	rm->atomic.op_rkey = rds_rdma_cookie_key(rm->m_rdma_cookie);
+	rm->atomic.op_rkey = rds_rdma_cookie_key(args->cookie);
 	rm->atomic.op_remote_addr = args->remote_addr + rds_rdma_cookie_offset(args->cookie);
 
 	rm->atomic.op_active = 1;