Merge tag 'mlx5-updates-2022-12-08' of git://git.kernel.org/pub/scm/l…

…inux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2022-12-08

1) Support range match action in SW steering

Yevgeny Kliteynik says:
=======================

The following patch series adds support for a range match action in
SW Steering.

SW steering is able to match only on the exact values of the packet fields,
as requested by the user: the user provides mask for the fields that are of
interest, and the exact values to be matched on when the traffic is handled.

The following patch series add new type of action - Range Match, where the
user provides a field to be matched on and a range of values (min to max)
that will be considered as hit.

There are several new notions that were implemented in order to support
Range Match:
 - MATCH_RANGES Steering Table Entry (STE): the new STE type that allows
   matching the packets' fields on the range of values instead of a specific
   value.
 - Match Definer: this is a general FW object that defines which fields
   in the packet will be referenced by the mask and tag of each STE.
   Match definer ID is part of STE fields, and it defines how the HW needs
   to interpret the STE's mask/tag values.
   Till now SW steering used the definers that were managed by FW and
   implemented the STE layout as described by the HW spec.
   Now that we're adding a new type of STE, SW steering needs to also be
   able to define this new STE's layout, and this is do

=======================

2) From OZ add support for meter mtu offload
   2.1: Refactor the code to allow both metering and range post actions as a
        pre-step for adding police mtu offload support.
   2.2: Instantiate mtu green/red flow tables with a single match-all rule.
        Add the green/red actions to the hit/miss table accordingly
   2.3: Initialize the meter object with the TC police mtu parameter.
        Use the hardware range match action feature.

3) From MaorD, support routes with more than 2 nexthops in multipath

4) Michael and Or, improve and extend vport representor counters.

* tag 'mlx5-updates-2022-12-08' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5: Expose steering dropped packets counter
  net/mlx5: Refactor and expand rep vport stat group
  net/mlx5e: multipath, support routes with more than 2 nexthops
  net/mlx5e: TC, add support for meter mtu offload
  net/mlx5e: meter, add mtu post meter tables
  net/mlx5e: meter, refactor to allow multiple post meter tables
  net/mlx5: DR, Add support for range match action
  net/mlx5: DR, Add function that tells if STE miss addr has been initialized
  net/mlx5: DR, Some refactoring of miss address handling
  net/mlx5: DR, Manage definers with refcounts
  net/mlx5: DR, Handle FT action in a separate function
  net/mlx5: DR, Rework is_fw_table function
  net/mlx5: DR, Add functions to create/destroy MATCH_DEFINER general object
  net/mlx5: fs, add match on ranges API
  net/mlx5: mlx5_ifc updates for MATCH_DEFINER general object
====================

Link: https://lore.kernel.org/r/20221209001420.142794-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/net/ethernet/mellanox/mlx5/core/Makefile

@@ -111,6 +111,7 @@ mlx5_core-$(CONFIG_MLX5_SW_STEERING) += steering/dr_domain.o steering/dr_table.o
 					steering/dr_ste_v2.o \
 					steering/dr_cmd.o steering/dr_fw.o \
 					steering/dr_action.o steering/fs_dr.o \
+					steering/dr_definer.o \
 					steering/dr_dbg.o lib/smfs.o
 #
 # SF device

drivers/net/ethernet/mellanox/mlx5/core/diag/fs_tracepoint.c

@@ -228,6 +228,17 @@ const char *parse_fs_hdrs(struct trace_seq *p,
 	return ret;
 }
 
+static const char
+*fs_dest_range_field_to_str(enum mlx5_flow_dest_range_field field)
+{
+	switch (field) {
+	case MLX5_FLOW_DEST_RANGE_FIELD_PKT_LEN:
+		return "packet len";
+	default:
+		return "unknown dest range field";
+	}
+}
+
 const char *parse_fs_dst(struct trace_seq *p,
 			 const struct mlx5_flow_destination *dst,
 			 u32 counter_id)
@@ -259,6 +270,11 @@ const char *parse_fs_dst(struct trace_seq *p,
 	case MLX5_FLOW_DESTINATION_TYPE_PORT:
 		trace_seq_printf(p, "port\n");
 		break;
+	case MLX5_FLOW_DESTINATION_TYPE_RANGE:
+		trace_seq_printf(p, "field=%s min=%d max=%d\n",
+				 fs_dest_range_field_to_str(dst->range.field),
+				 dst->range.min, dst->range.max);
+		break;
 	case MLX5_FLOW_DESTINATION_TYPE_NONE:
 		trace_seq_printf(p, "none\n");
 		break;

drivers/net/ethernet/mellanox/mlx5/core/en/tc/act/police.c

@@ -3,6 +3,7 @@
 
 #include "act.h"
 #include "en/tc_priv.h"
+#include "fs_core.h"
 
 static bool police_act_validate_control(enum flow_action_id act_id,
 					struct netlink_ext_ack *extack)
@@ -71,6 +72,8 @@ fill_meter_params_from_act(const struct flow_action_entry *act,
 		params->mode = MLX5_RATE_LIMIT_PPS;
 		params->rate = act->police.rate_pkt_ps;
 		params->burst = act->police.burst_pkt;
+	} else if (act->police.mtu) {
+		params->mtu = act->police.mtu;
 	} else {
 		return -EOPNOTSUPP;
 	}
@@ -84,14 +87,25 @@ tc_act_parse_police(struct mlx5e_tc_act_parse_state *parse_state,
 		    struct mlx5e_priv *priv,
 		    struct mlx5_flow_attr *attr)
 {
+	enum mlx5_flow_namespace_type ns =  mlx5e_get_flow_namespace(parse_state->flow);
+	struct mlx5e_flow_meter_params *params = &attr->meter_attr.params;
 	int err;
 
-	err = fill_meter_params_from_act(act, &attr->meter_attr.params);
+	err = fill_meter_params_from_act(act, params);
 	if (err)
 		return err;
 
-	attr->action |= MLX5_FLOW_CONTEXT_ACTION_EXECUTE_ASO;
-	attr->exe_aso_type = MLX5_EXE_ASO_FLOW_METER;
+	if (params->mtu) {
+		if (!(mlx5_fs_get_capabilities(priv->mdev, ns) &
+		      MLX5_FLOW_STEERING_CAP_MATCH_RANGES))
+			return -EOPNOTSUPP;
+
+		attr->action |= MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
+		attr->flags |= MLX5_ATTR_FLAG_MTU;
+	} else {
+		attr->action |= MLX5_FLOW_CONTEXT_ACTION_EXECUTE_ASO;
+		attr->exe_aso_type = MLX5_EXE_ASO_FLOW_METER;
+	}
 
 	return 0;
 }

drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.c

@@ -241,7 +241,7 @@ mlx5e_flow_meter_destroy_aso_obj(struct mlx5_core_dev *mdev, u32 obj_id)
 }
 
 static struct mlx5e_flow_meter_handle *
-__mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters)
+__mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters, bool alloc_aso)
 {
 	struct mlx5_core_dev *mdev = flow_meters->mdev;
 	struct mlx5e_flow_meter_aso_obj *meters_obj;
@@ -268,6 +268,9 @@ __mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters)
 	}
 	meter->act_counter = counter;
 
+	if (!alloc_aso)
+		goto no_aso;
+
 	meters_obj = list_first_entry_or_null(&flow_meters->partial_list,
 					      struct mlx5e_flow_meter_aso_obj,
 					      entry);
@@ -300,11 +303,12 @@ __mlx5e_flow_meter_alloc(struct mlx5e_flow_meters *flow_meters)
 	}
 
 	bitmap_set(meters_obj->meters_map, pos, 1);
-	meter->flow_meters = flow_meters;
 	meter->meters_obj = meters_obj;
 	meter->obj_id = meters_obj->base_id + pos / 2;
 	meter->idx = pos % 2;
 
+no_aso:
+	meter->flow_meters = flow_meters;
 	mlx5_core_dbg(mdev, "flow meter allocated, obj_id=0x%x, index=%d\n",
 		      meter->obj_id, meter->idx);
 
@@ -332,6 +336,9 @@ __mlx5e_flow_meter_free(struct mlx5e_flow_meter_handle *meter)
 	mlx5_fc_destroy(mdev, meter->act_counter);
 	mlx5_fc_destroy(mdev, meter->drop_counter);
 
+	if (meter->params.mtu)
+		goto out_no_aso;
+
 	meters_obj = meter->meters_obj;
 	pos = (meter->obj_id - meters_obj->base_id) * 2 + meter->idx;
 	bitmap_clear(meters_obj->meters_map, pos, 1);
@@ -345,6 +352,7 @@ __mlx5e_flow_meter_free(struct mlx5e_flow_meter_handle *meter)
 		list_add(&meters_obj->entry, &flow_meters->partial_list);
 	}
 
+out_no_aso:
 	mlx5_core_dbg(mdev, "flow meter freed, obj_id=0x%x, index=%d\n",
 		      meter->obj_id, meter->idx);
 	kfree(meter);
@@ -409,12 +417,13 @@ mlx5e_tc_meter_alloc(struct mlx5e_flow_meters *flow_meters,
 {
 	struct mlx5e_flow_meter_handle *meter;
 
-	meter = __mlx5e_flow_meter_alloc(flow_meters);
+	meter = __mlx5e_flow_meter_alloc(flow_meters, !params->mtu);
 	if (IS_ERR(meter))
 		return meter;
 
 	hash_add(flow_meters->hashtbl, &meter->hlist, params->index);
 	meter->params.index = params->index;
+	meter->params.mtu = params->mtu;
 	meter->refcnt++;
 
 	return meter;

drivers/net/ethernet/mellanox/mlx5/core/en/tc/meter.h

@@ -20,6 +20,7 @@ struct mlx5e_flow_meter_params {
 	u32 index;
 	u64 rate;
 	u64 burst;
+	u32 mtu;
 };
 
 struct mlx5e_flow_meter_handle {