IB/mlx5: Support congestion related counters

PATCH Applied Leon Romanovsky 2 participants 2 messages Apr 16, 2017 – Apr 24, 2017 1 message not in the archive

Subsystems (3)

InfiniBand Mellanox mlx5 core Mellanox MLX5 IB driver

lore

Share & tools

lore

Leon Romanovsky

author

2017-04-16 04:29 UTC #0 lore raw

PATCH v2 1/1 IB/mlx5: Support congestion related counters

From: Parav Pandit <parav-VPRAkNaXOzVWk0Htik3J/[email protected]>

This patch adds support to query the congestion related hardware counters through new command and links them with other hw counters being available in hw_counters sysfs location.

In order to reuse existing infrastructure it renames related q_counter data structures to more generic counters to reflect q_counters and congestion counters and maybe some other counters in the future.

New hardware counters:
* rp_cnp_handled - CNP packets handled by the reaction point
* rp_cnp_ignored - CNP packets ignored by the reaction point
* np_cnp_sent - CNP packets sent by notification point to respond to
CE marked RoCE packets
* np_ecn_marked_roce_packets - CE marked RoCE packets received by notification point

It also avoids returning ENOSYS which is specific for invalid system call and produces the following checkpatch.pl warning.

WARNING: ENOSYS means 'invalid syscall nr' and nothing else
+ return -ENOSYS;

Signed-off-by Parav Pandit <parav-VPRAkNaXOzVWk0Htik3J/[email protected]>

Reviewed-by Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/[email protected]>

Reviewed-by Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/[email protected]>

Signed-off-by Leon Romanovsky <[email protected]>

---
Changes from v1:
* Fix kbuild error, which was result of me sending wrong patch.
* Convert all congestion counters (*_cnp_*) to be in singular form, like most of other counters (Q counters).
Changes from v0:
* Fix overflow issue in access to congestion counters

---
 drivers/infiniband/hw/mlx5/cmd.c     |  11 +++
 drivers/infiniband/hw/mlx5/cmd.h     |   2 +
 drivers/infiniband/hw/mlx5/main.c    | 172 ++++++++++++++++++++++++-----------
 drivers/infiniband/hw/mlx5/mlx5_ib.h |   7 +-
 drivers/infiniband/hw/mlx5/qp.c      |   7 +-
 include/linux/mlx5/mlx5_ifc.h        |  18 ++--
 6 files changed, 150 insertions(+), 67 deletions(-)

drivers/infiniband/hw/mlx5/cmd.c +11 -0

index cdc2d3017da7..18d5e1db93ed 100644
@@ -46,3 +46,14 @@ int mlx5_cmd_null_mkey(struct mlx5_core_dev *dev, u32 *null_mkey)
 				      null_mkey);
 	return err;
 }
+
+int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
+				bool reset, void *out, int out_size)
+{
+	u32 in[MLX5_ST_SZ_DW(query_cong_statistics_in)] = { };
+
+	MLX5_SET(query_cong_statistics_in, in, opcode,
+		 MLX5_CMD_OP_QUERY_CONG_STATISTICS);
+	MLX5_SET(query_cong_statistics_in, in, clear, reset);
+	return mlx5_cmd_exec(dev, in, sizeof(in), out, out_size);
+}

drivers/infiniband/hw/mlx5/cmd.h +2 -0

index 7ca8a7b6434d..fa09228193a6 100644
@@ -37,4 +37,6 @@
 #include <linux/mlx5/driver.h>
 
 int mlx5_cmd_null_mkey(struct mlx5_core_dev *dev, u32 *null_mkey);
+int mlx5_cmd_query_cong_counter(struct mlx5_core_dev *dev,
+				bool reset, void *out, int out_size);
 #endif /* MLX5_IB_CMD_H */

drivers/infiniband/hw/mlx5/main.c +120 ~~-52~~

312 lines not on this page · raw message

drivers/infiniband/hw/mlx5/mlx5_ib.h +4 -3

index 3cd064b5f0bf..7f1221cb6a50 100644
@@ -595,15 +595,16 @@ struct mlx5_ib_resources {
 	struct mutex	mutex;
 };
 
-struct mlx5_ib_q_counters {
+struct mlx5_ib_counters {
 	const char **names;
 	size_t *offsets;
-	u32 num_counters;
+	u32 num_q_counters;
+	u32 num_cong_counters;
 	u16 set_id;
 };
 
 struct mlx5_ib_port {
-	struct mlx5_ib_q_counters q_cnts;
+	struct mlx5_ib_counters cnts;
 };
 
 struct mlx5_roce {

drivers/infiniband/hw/mlx5/qp.c +4 -3

index ad8a2638e339..74a4696b4833 100644
@@ -2798,7 +2798,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 			       qp->port) - 1;
 		mibport = &dev->port[port_num];
 		context->qp_counter_set_usr_page |=
-			cpu_to_be32((u32)(mibport->q_cnts.set_id) << 24);
+			cpu_to_be32((u32)(mibport->cnts.set_id) << 24);
 	}
 
 	if (!ibqp->uobject && cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT)

@@ -2826,7 +2826,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 
 		raw_qp_param.operation = op;
 		if (cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT) {
-			raw_qp_param.rq_q_ctr_id = mibport->q_cnts.set_id;
+			raw_qp_param.rq_q_ctr_id = mibport->cnts.set_id;
 			raw_qp_param.set_mask |= MLX5_RAW_QP_MOD_SET_RQ_Q_CTR_ID;
 		}

@@ -4964,7 +4964,8 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
 		if (MLX5_CAP_GEN(dev->mdev, modify_rq_counter_set_id)) {
 			MLX5_SET64(modify_rq_in, in, modify_bitmask,
 				   MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_RQ_COUNTER_SET_ID);
-			MLX5_SET(rqc, rqc, counter_set_id, dev->port->q_cnts.set_id);
+			MLX5_SET(rqc, rqc, counter_set_id,
+				 dev->port->cnts.set_id);
 		} else
 			pr_info_once("%s: Receive WQ counters are not supported on current FW\n",
 				     dev->ib_dev.name);

include/linux/mlx5/mlx5_ifc.h +9 -9

index 838242697541..8ecfdd057ae9 100644
@@ -4623,17 +4623,17 @@ struct mlx5_ifc_query_cong_statistics_out_bits {
 
 	u8         reserved_at_40[0x40];
 
-	u8         cur_flows[0x20];
+	u8         rp_cur_flows[0x20];
 
 	u8         sum_flows[0x20];
 
-	u8         cnp_ignored_high[0x20];
+	u8         rp_cnp_ignored_high[0x20];
 
-	u8         cnp_ignored_low[0x20];
+	u8         rp_cnp_ignored_low[0x20];
 
-	u8         cnp_handled_high[0x20];
+	u8         rp_cnp_handled_high[0x20];
 
-	u8         cnp_handled_low[0x20];
+	u8         rp_cnp_handled_low[0x20];
 
 	u8         reserved_at_140[0x100];

@@ -4643,13 +4643,13 @@ struct mlx5_ifc_query_cong_statistics_out_bits {
 
 	u8         accumulators_period[0x20];
 
-	u8         ecn_marked_roce_packets_high[0x20];
+	u8         np_ecn_marked_roce_packets_high[0x20];
 
-	u8         ecn_marked_roce_packets_low[0x20];
+	u8         np_ecn_marked_roce_packets_low[0x20];
 
-	u8         cnps_sent_high[0x20];
+	u8         np_cnp_sent_high[0x20];
 
-	u8         cnps_sent_low[0x20];
+	u8         np_cnp_sent_low[0x20];
 
 	u8         reserved_at_320[0x560];
 };

--
2.12.2

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

A message that is not in the archive (sent off-list or not yet received): 20170416042929.18294-1-leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org

Doug Ledford

Applied

2017-04-24 16:26 UTC #2 lore raw

Quoted 25 lines · Leon Romanovsky

From: Parav Pandit <parav-VPRAkNaXOzVWk0Htik3J/[email protected]>
This patch adds support to query the congestion related hardware counters
through new command and links them with other hw counters being available
in hw_counters sysfs location.
In order to reuse existing infrastructure it renames related q_counter
data structures to more generic counters to reflect q_counters and congestion counters and maybe some other counters in the future.
New hardware counters:
* rp_cnp_handled - CNP packets handled by the reaction point
* rp_cnp_ignored - CNP packets ignored by the reaction point
* np_cnp_sent - CNP packets sent by notification point to respond to
CE marked RoCE packets
* np_ecn_marked_roce_packets - CE marked RoCE packets received by notification point
It also avoids returning ENOSYS which is specific for invalid system call and produces the following checkpatch.pl warning.

Quoted 7 lines

WARNING: ENOSYS means 'invalid syscall nr' and nothing else + return -ENOSYS;
Signed-off-by Parav Pandit <parav-VPRAkNaXOzVWk0Htik3J/[email protected]>
Reviewed-by Eli Cohen <eli-VPRAkNaXOzVWk0Htik3J/[email protected]>
Reviewed-by Daniel Jurgens <danielj-VPRAkNaXOzVWk0Htik3J/[email protected]>
Signed-off-by Leon Romanovsky <[email protected]>

Thanks, applied.

Signature · 9 lines

Doug Ledford <[email protected]>
    GPG KeyID: B826A3330E572FDD
   
Key fingerprint = AE6B 1BDA 122B 23B4 265B  1274 B826 A333 0E57 2FDD

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html