Merge branch 'qdisc-RED-offload'

Jiri Pirko says:

====================
qdisc RED offload

Nogah says:

Add an offload support for RED qdisc for mlxsw driver.
The first patch adds the ability to offload RED qdisc by using
ndo_setup_tc. It gives RED three commands, to offload, change or delete
the qdisc, to get the qdisc generic stats and to get it's RED xstats.
There is no enforcement on a driver to offload or not offload the qdisc and
it is up to the driver to decide.
RED qdisc is first being created and only later graft to a parent (unless
it is a root qdisc). For that reason the return value of the offload
replace command that is called in the init process doesn't reflect actual
offload state. The offload state is determined in the dump function so it
can be reflected to the user. This function is also responsible for stats
update.

The patchses 2-3 change the name of TC_SETUP_MQPRIO & TC_SETUP_CBS to match
with the new convention of QDISC prefix.
The rest of the patchset is driver support for the qdisc. Currently only
as root qdisc that is being set on the default traffic class. It supports
only the following parameters of RED: min, max, probability and ECN mode.
Limit and burst size related params are being ignored at this moment.

---
v7->v8 internal: (external RFC->v1)
- patch 1/9:
- unite the offload and un-offload functions
- clean the OFFLOAD flag when the qdisc in not offloaded
- patch 2/9:
- minor change to avoid a conflict
- patch 5/9:
- check for bad min/max values
- clean the offloaded qdisc after a bad config call
====================

Signed-off-by: David S. Miller <[e-mail]>

drivers/net/ethernet/amd/xgbe/xgbe-drv.c

···
 	struct tc_mqprio_qopt *mqprio = type_data;
 	u8 tc;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c

···
 {
 	struct tc_mqprio_qopt *mqprio = type_data;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

drivers/net/ethernet/broadcom/bnxt/bnxt.c

···
 	switch (type) {
 	case TC_SETUP_BLOCK:
 		return bnxt_setup_tc_block(dev, type_data);
-	case TC_SETUP_MQPRIO: {
+	case TC_SETUP_QDISC_MQPRIO: {
 		struct tc_mqprio_qopt *mqprio = type_data;
 
 		mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

drivers/net/ethernet/freescale/dpaa/dpaa_eth.c

···
 	u8 num_tc;
 	int i;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

drivers/net/ethernet/hisilicon/hns3/hns3pf/hns3_enet.c

···
 static int hns3_nic_setup_tc(struct net_device *dev, enum tc_setup_type type,
 			     void *type_data)
 {
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	return hns3_setup_tc(dev, type_data);

drivers/net/ethernet/intel/fm10k/fm10k_netdev.c

···
 {
 	struct tc_mqprio_qopt *mqprio = type_data;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

drivers/net/ethernet/intel/i40e/i40e_main.c

···
 			   void *type_data)
 {
 	switch (type) {
-	case TC_SETUP_MQPRIO:
+	case TC_SETUP_QDISC_MQPRIO:
 		return i40e_setup_tc(netdev, type_data);
 	case TC_SETUP_BLOCK:
 		return i40e_setup_tc_block(netdev, type_data);

drivers/net/ethernet/intel/igb/igb_main.c

···
 	struct igb_adapter *adapter = netdev_priv(dev);
 
 	switch (type) {
-	case TC_SETUP_CBS:
+	case TC_SETUP_QDISC_CBS:
 		return igb_offload_cbs(adapter, type_data);
 
 	default:

drivers/net/ethernet/intel/ixgbe/ixgbe_main.c

···
 	switch (type) {
 	case TC_SETUP_BLOCK:
 		return ixgbe_setup_tc_block(dev, type_data);
-	case TC_SETUP_MQPRIO:
+	case TC_SETUP_QDISC_MQPRIO:
 		return ixgbe_setup_tc_mqprio(dev, type_data);
 	default:
 		return -EOPNOTSUPP;

drivers/net/ethernet/mellanox/mlx4/en_netdev.c

···
 {
 	struct tc_mqprio_qopt *mqprio = type_data;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	if (mqprio->num_tc && mqprio->num_tc != MLX4_EN_NUM_UP_HIGH)

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

···
 	case TC_SETUP_BLOCK:
 		return mlx5e_setup_tc_block(dev, type_data);
 #endif
-	case TC_SETUP_MQPRIO:
+	case TC_SETUP_QDISC_MQPRIO:
 		return mlx5e_setup_tc_mqprio(dev, type_data);
 	default:
 		return -EOPNOTSUPP;

drivers/net/ethernet/mellanox/mlxsw/Makefile

···
 				   spectrum_acl.o spectrum_flower.o \
 				   spectrum_cnt.o spectrum_fid.o \
 				   spectrum_ipip.o spectrum_acl_flex_actions.o \
-				   spectrum_mr.o spectrum_mr_tcam.o
+				   spectrum_mr.o spectrum_mr_tcam.o \
+				   spectrum_qdisc.o
 mlxsw_spectrum-$(CONFIG_MLXSW_SPECTRUM_DCB)	+= spectrum_dcb.o
 mlxsw_spectrum-$(CONFIG_NET_DEVLINK) += spectrum_dpipe.o
 obj-$(CONFIG_MLXSW_MINIMAL)	+= mlxsw_minimal.o

drivers/net/ethernet/mellanox/mlxsw/reg.h

···
 	}
 }
 
+/* CWTP - Congetion WRED ECN TClass Profile
+ * ----------------------------------------
+ * Configures the profiles for queues of egress port and traffic class
+ */
+#define MLXSW_REG_CWTP_ID 0x2802
+#define MLXSW_REG_CWTP_BASE_LEN 0x28
+#define MLXSW_REG_CWTP_PROFILE_DATA_REC_LEN 0x08
+#define MLXSW_REG_CWTP_LEN 0x40
+
+MLXSW_REG_DEFINE(cwtp, MLXSW_REG_CWTP_ID, MLXSW_REG_CWTP_LEN);
+
+/* reg_cwtp_local_port
+ * Local port number
+ * Not supported for CPU port
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, cwtp, local_port, 0, 16, 8);
+
+/* reg_cwtp_traffic_class
+ * Traffic Class to configure
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, cwtp, traffic_class, 32, 0, 8);
+
+/* reg_cwtp_profile_min
+ * Minimum Average Queue Size of the profile in cells.
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, cwtp, profile_min, MLXSW_REG_CWTP_BASE_LEN,
+		     0, 20, MLXSW_REG_CWTP_PROFILE_DATA_REC_LEN, 0, false);
+
+/* reg_cwtp_profile_percent
+ * Percentage of WRED and ECN marking for maximum Average Queue size
+ * Range is 0 to 100, units of integer percentage
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, cwtp, profile_percent, MLXSW_REG_CWTP_BASE_LEN,
+		     24, 7, MLXSW_REG_CWTP_PROFILE_DATA_REC_LEN, 4, false);
+
+/* reg_cwtp_profile_max
+ * Maximum Average Queue size of the profile in cells
+ * Access: RW
+ */
+MLXSW_ITEM32_INDEXED(reg, cwtp, profile_max, MLXSW_REG_CWTP_BASE_LEN,
+		     0, 20, MLXSW_REG_CWTP_PROFILE_DATA_REC_LEN, 4, false);
+
+#define MLXSW_REG_CWTP_MIN_VALUE 64
+#define MLXSW_REG_CWTP_MAX_PROFILE 2
+#define MLXSW_REG_CWTP_DEFAULT_PROFILE 1
+
+static inline void mlxsw_reg_cwtp_pack(char *payload, u8 local_port,
+				       u8 traffic_class)
+{
+	int i;
+
+	MLXSW_REG_ZERO(cwtp, payload);
+	mlxsw_reg_cwtp_local_port_set(payload, local_port);
+	mlxsw_reg_cwtp_traffic_class_set(payload, traffic_class);
+
+	for (i = 0; i <= MLXSW_REG_CWTP_MAX_PROFILE; i++) {
+		mlxsw_reg_cwtp_profile_min_set(payload, i,
+					       MLXSW_REG_CWTP_MIN_VALUE);
+		mlxsw_reg_cwtp_profile_max_set(payload, i,
+					       MLXSW_REG_CWTP_MIN_VALUE);
+	}
+}
+
+#define MLXSW_REG_CWTP_PROFILE_TO_INDEX(profile) (profile - 1)
+
+static inline void
+mlxsw_reg_cwtp_profile_pack(char *payload, u8 profile, u32 min, u32 max,
+			    u32 probability)
+{
+	u8 index = MLXSW_REG_CWTP_PROFILE_TO_INDEX(profile);
+
+	mlxsw_reg_cwtp_profile_min_set(payload, index, min);
+	mlxsw_reg_cwtp_profile_max_set(payload, index, max);
+	mlxsw_reg_cwtp_profile_percent_set(payload, index, probability);
+}
+
+/* CWTPM - Congestion WRED ECN TClass and Pool Mapping
+ * ---------------------------------------------------
+ * The CWTPM register maps each egress port and traffic class to profile num.
+ */
+#define MLXSW_REG_CWTPM_ID 0x2803
+#define MLXSW_REG_CWTPM_LEN 0x44
+
+MLXSW_REG_DEFINE(cwtpm, MLXSW_REG_CWTPM_ID, MLXSW_REG_CWTPM_LEN);
+
+/* reg_cwtpm_local_port
+ * Local port number
+ * Not supported for CPU port
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, cwtpm, local_port, 0, 16, 8);
+
+/* reg_cwtpm_traffic_class
+ * Traffic Class to configure
+ * Access: Index
+ */
+MLXSW_ITEM32(reg, cwtpm, traffic_class, 32, 0, 8);
+
+/* reg_cwtpm_ew
+ * Control enablement of WRED for traffic class:
+ * 0 - Disable
+ * 1 - Enable
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, ew, 36, 1, 1);
+
+/* reg_cwtpm_ee
+ * Control enablement of ECN for traffic class:
+ * 0 - Disable
+ * 1 - Enable
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, ee, 36, 0, 1);
+
+/* reg_cwtpm_tcp_g
+ * TCP Green Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, tcp_g, 52, 0, 2);
+
+/* reg_cwtpm_tcp_y
+ * TCP Yellow Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, tcp_y, 56, 16, 2);
+
+/* reg_cwtpm_tcp_r
+ * TCP Red Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, tcp_r, 56, 0, 2);
+
+/* reg_cwtpm_ntcp_g
+ * Non-TCP Green Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, ntcp_g, 60, 0, 2);
+
+/* reg_cwtpm_ntcp_y
+ * Non-TCP Yellow Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, ntcp_y, 64, 16, 2);
+
+/* reg_cwtpm_ntcp_r
+ * Non-TCP Red Profile.
+ * Index of the profile within {port, traffic class} to use.
+ * 0 for disabling both WRED and ECN for this type of traffic.
+ * Access: RW
+ */
+MLXSW_ITEM32(reg, cwtpm, ntcp_r, 64, 0, 2);
+
+#define MLXSW_REG_CWTPM_RESET_PROFILE 0
+
+static inline void mlxsw_reg_cwtpm_pack(char *payload, u8 local_port,
+					u8 traffic_class, u8 profile,
+					bool wred, bool ecn)
+{
+	MLXSW_REG_ZERO(cwtpm, payload);
+	mlxsw_reg_cwtpm_local_port_set(payload, local_port);
+	mlxsw_reg_cwtpm_traffic_class_set(payload, traffic_class);
+	mlxsw_reg_cwtpm_ew_set(payload, wred);
+	mlxsw_reg_cwtpm_ee_set(payload, ecn);
+	mlxsw_reg_cwtpm_tcp_g_set(payload, profile);
+	mlxsw_reg_cwtpm_tcp_y_set(payload, profile);
+	mlxsw_reg_cwtpm_tcp_r_set(payload, profile);
+	mlxsw_reg_cwtpm_ntcp_g_set(payload, profile);
+	mlxsw_reg_cwtpm_ntcp_y_set(payload, profile);
+	mlxsw_reg_cwtpm_ntcp_r_set(payload, profile);
+}
+
 /* PPBT - Policy-Engine Port Binding Table
  * ---------------------------------------
  * This register is used for configuration of the Port Binding Table.
···
 
 enum mlxsw_reg_ppcnt_grp {
 	MLXSW_REG_PPCNT_IEEE_8023_CNT = 0x0,
+	MLXSW_REG_PPCNT_EXT_CNT = 0x5,
 	MLXSW_REG_PPCNT_PRIO_CNT = 0x10,
 	MLXSW_REG_PPCNT_TC_CNT = 0x11,
+	MLXSW_REG_PPCNT_TC_CONG_TC = 0x13,
 };
 
 /* reg_ppcnt_grp
···
  * 0x10: Per Priority Counters
  * 0x11: Per Traffic Class Counters
  * 0x12: Physical Layer Counters
+ * 0x13: Per Traffic Class Congestion Counters
  * Access: Index
  */
 MLXSW_ITEM32(reg, ppcnt, grp, 0x00, 0, 6);
···
 MLXSW_ITEM64(reg, ppcnt, a_pause_mac_ctrl_frames_transmitted,
 	     MLXSW_REG_PPCNT_COUNTERS_OFFSET + 0x90, 0, 64);
 
+/* Ethernet Extended Counter Group Counters */
+
+/* reg_ppcnt_ecn_marked
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, ecn_marked,
+	     MLXSW_REG_PPCNT_COUNTERS_OFFSET + 0x08, 0, 64);
+
 /* Ethernet Per Priority Group Counters */
 
 /* reg_ppcnt_rx_octets
···
  */
 MLXSW_ITEM64(reg, ppcnt, tc_no_buffer_discard_uc,
 	     MLXSW_REG_PPCNT_COUNTERS_OFFSET + 0x08, 0, 64);
+
+/* Ethernet Per Traffic Class Congestion Group Counters */
+
+/* reg_ppcnt_wred_discard
+ * Access: RO
+ */
+MLXSW_ITEM64(reg, ppcnt, wred_discard,
+	     MLXSW_REG_PPCNT_COUNTERS_OFFSET + 0x00, 0, 64);
 
 static inline void mlxsw_reg_ppcnt_pack(char *payload, u8 local_port,
 					enum mlxsw_reg_ppcnt_grp grp,
···
 	MLXSW_REG(svpe),
 	MLXSW_REG(sfmr),
 	MLXSW_REG(spvmlr),
+	MLXSW_REG(cwtp),
+	MLXSW_REG(cwtpm),
 	MLXSW_REG(ppbt),
 	MLXSW_REG(pacl),
 	MLXSW_REG(pagt),

drivers/net/ethernet/mellanox/mlxsw/spectrum.c

···
 	return err;
 }
 
+static void
+mlxsw_sp_port_get_hw_xstats(struct net_device *dev,
+			    struct mlxsw_sp_port_xstats *xstats)
+{
+	char ppcnt_pl[MLXSW_REG_PPCNT_LEN];
+	int err, i;
+
+	err = mlxsw_sp_port_get_stats_raw(dev, MLXSW_REG_PPCNT_EXT_CNT, 0,
+					  ppcnt_pl);
+	if (!err)
+		xstats->ecn = mlxsw_reg_ppcnt_ecn_marked_get(ppcnt_pl);
+
+	for (i = 0; i < TC_MAX_QUEUE; i++) {
+		err = mlxsw_sp_port_get_stats_raw(dev,
+						  MLXSW_REG_PPCNT_TC_CONG_TC,
+						  i, ppcnt_pl);
+		if (!err)
+			xstats->wred_drop[i] =
+				mlxsw_reg_ppcnt_wred_discard_get(ppcnt_pl);
+
+		err = mlxsw_sp_port_get_stats_raw(dev, MLXSW_REG_PPCNT_TC_CNT,
+						  i, ppcnt_pl);
+		if (err)
+			continue;
+
+		xstats->backlog[i] =
+			mlxsw_reg_ppcnt_tc_transmit_queue_get(ppcnt_pl);
+		xstats->tail_drop[i] =
+			mlxsw_reg_ppcnt_tc_no_buffer_discard_uc_get(ppcnt_pl);
+	}
+}
+
 static void update_stats_cache(struct work_struct *work)
 {
 	struct mlxsw_sp_port *mlxsw_sp_port =
···
 
 	mlxsw_sp_port_get_hw_stats(mlxsw_sp_port->dev,
 				   &mlxsw_sp_port->periodic_hw_stats.stats);
+	mlxsw_sp_port_get_hw_xstats(mlxsw_sp_port->dev,
+				    &mlxsw_sp_port->periodic_hw_stats.xstats);
 
 out:
 	mlxsw_core_schedule_dw(&mlxsw_sp_port->periodic_hw_stats.update_dw,
···
 	switch (type) {
 	case TC_SETUP_BLOCK:
 		return mlxsw_sp_setup_tc_block(mlxsw_sp_port, type_data);
+	case TC_SETUP_QDISC_RED:
+		return mlxsw_sp_setup_tc_red(mlxsw_sp_port, type_data);
 	default:
 		return -EOPNOTSUPP;
 	}

drivers/net/ethernet/mellanox/mlxsw/spectrum.h

···
 #include <linux/notifier.h>
 #include <net/psample.h>
 #include <net/pkt_cls.h>
+#include <net/red.h>
 
 #include "port.h"
 #include "core.h"
···
 	struct list_head bridge_vlan_node;
 };
 
+enum mlxsw_sp_qdisc_type {
+	MLXSW_SP_QDISC_NO_QDISC,
+	MLXSW_SP_QDISC_RED,
+};
+
+struct mlxsw_sp_qdisc {
+	u32 handle;
+	enum mlxsw_sp_qdisc_type type;
+	struct red_stats xstats_base;
+	union {
+		struct {
+			u64 tail_drop_base;
+			u64 ecn_base;
+			u64 wred_drop_base;
+		} red;
+	} xstats;
+
+	u64 tx_bytes;
+	u64 tx_packets;
+	u64 drops;
+	u64 overlimits;
+};
+
+/* No need an internal lock; At worse - miss a single periodic iteration */
+struct mlxsw_sp_port_xstats {
+	u64 ecn;
+	u64 wred_drop[TC_MAX_QUEUE];
+	u64 tail_drop[TC_MAX_QUEUE];
+	u64 backlog[TC_MAX_QUEUE];
+};
+
 struct mlxsw_sp_port {
 	struct net_device *dev;
 	struct mlxsw_sp_port_pcpu_stats __percpu *pcpu_stats;
···
 	struct {
 		#define MLXSW_HW_STATS_UPDATE_TIME HZ
 		struct rtnl_link_stats64 stats;
+		struct mlxsw_sp_port_xstats xstats;
 		struct delayed_work update_dw;
 	} periodic_hw_stats;
 	struct mlxsw_sp_port_sample *sample;
 	struct list_head vlans_list;
+	struct mlxsw_sp_qdisc root_qdisc;
 };
 
 static inline bool
···
 			     struct tc_cls_flower_offload *f);
 int mlxsw_sp_flower_stats(struct mlxsw_sp_port *mlxsw_sp_port, bool ingress,
 			  struct tc_cls_flower_offload *f);
+
+/* spectrum_qdisc.c */
+int mlxsw_sp_setup_tc_red(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_red_qopt_offload *p);
 
 /* spectrum_fid.c */
 int mlxsw_sp_fid_flood_set(struct mlxsw_sp_fid *fid,

drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c

···
+/*
+ * drivers/net/ethernet/mellanox/mlxsw/spectrum_qdisc.c
+ * Copyright (c) 2017 Mellanox Technologies. All rights reserved.
+ * Copyright (c) 2017 Nogah Frankel <[e-mail]>
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. Neither the names of the copyright holders nor the names of its
+ *    contributors may be used to endorse or promote products derived from
+ *    this software without specific prior written permission.
+ *
+ * Alternatively, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") version 2 as published by the Free
+ * Software Foundation.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/netdevice.h>
+#include <net/pkt_cls.h>
+#include <net/red.h>
+
+#include "spectrum.h"
+#include "reg.h"
+
+static int
+mlxsw_sp_tclass_congestion_enable(struct mlxsw_sp_port *mlxsw_sp_port,
+				  int tclass_num, u32 min, u32 max,
+				  u32 probability, bool is_ecn)
+{
+	char cwtp_cmd[max_t(u8, MLXSW_REG_CWTP_LEN, MLXSW_REG_CWTPM_LEN)];
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	int err;
+
+	mlxsw_reg_cwtp_pack(cwtp_cmd, mlxsw_sp_port->local_port, tclass_num);
+	mlxsw_reg_cwtp_profile_pack(cwtp_cmd, MLXSW_REG_CWTP_DEFAULT_PROFILE,
+				    roundup(min, MLXSW_REG_CWTP_MIN_VALUE),
+				    roundup(max, MLXSW_REG_CWTP_MIN_VALUE),
+				    probability);
+
+	err = mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(cwtp), cwtp_cmd);
+	if (err)
+		return err;
+
+	mlxsw_reg_cwtpm_pack(cwtp_cmd, mlxsw_sp_port->local_port, tclass_num,
+			     MLXSW_REG_CWTP_DEFAULT_PROFILE, true, is_ecn);
+
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(cwtpm), cwtp_cmd);
+}
+
+static int
+mlxsw_sp_tclass_congestion_disable(struct mlxsw_sp_port *mlxsw_sp_port,
+				   int tclass_num)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	char cwtpm_cmd[MLXSW_REG_CWTPM_LEN];
+
+	mlxsw_reg_cwtpm_pack(cwtpm_cmd, mlxsw_sp_port->local_port, tclass_num,
+			     MLXSW_REG_CWTPM_RESET_PROFILE, false, false);
+	return mlxsw_reg_write(mlxsw_sp->core, MLXSW_REG(cwtpm), cwtpm_cmd);
+}
+
+static void
+mlxsw_sp_setup_tc_qdisc_clean_stats(struct mlxsw_sp_port *mlxsw_sp_port,
+				    struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+				    int tclass_num)
+{
+	struct red_stats *xstats_base = &mlxsw_sp_qdisc->xstats_base;
+	struct mlxsw_sp_port_xstats *xstats;
+	struct rtnl_link_stats64 *stats;
+
+	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
+	stats = &mlxsw_sp_port->periodic_hw_stats.stats;
+
+	mlxsw_sp_qdisc->tx_packets = stats->tx_packets;
+	mlxsw_sp_qdisc->tx_bytes = stats->tx_bytes;
+
+	switch (mlxsw_sp_qdisc->type) {
+	case MLXSW_SP_QDISC_RED:
+		xstats_base->prob_mark = xstats->ecn;
+		xstats_base->prob_drop = xstats->wred_drop[tclass_num];
+		xstats_base->pdrop = xstats->tail_drop[tclass_num];
+
+		mlxsw_sp_qdisc->overlimits = xstats_base->prob_drop +
+					     xstats_base->prob_mark;
+		mlxsw_sp_qdisc->drops = xstats_base->prob_drop +
+					xstats_base->pdrop;
+		break;
+	default:
+		break;
+	}
+}
+
+static int
+mlxsw_sp_qdisc_red_destroy(struct mlxsw_sp_port *mlxsw_sp_port, u32 handle,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			   int tclass_num)
+{
+	int err;
+
+	if (mlxsw_sp_qdisc->handle != handle)
+		return 0;
+
+	err = mlxsw_sp_tclass_congestion_disable(mlxsw_sp_port, tclass_num);
+	mlxsw_sp_qdisc->handle = TC_H_UNSPEC;
+	mlxsw_sp_qdisc->type = MLXSW_SP_QDISC_NO_QDISC;
+
+	return err;
+}
+
+static int
+mlxsw_sp_qdisc_red_replace(struct mlxsw_sp_port *mlxsw_sp_port, u32 handle,
+			   struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			   int tclass_num,
+			   struct tc_red_qopt_offload_params *p)
+{
+	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_port->mlxsw_sp;
+	u32 min, max;
+	u64 prob;
+	int err = 0;
+
+	if (p->min > p->max) {
+		dev_err(mlxsw_sp->bus_info->dev,
+			"spectrum: RED: min %u is bigger then max %u\n", p->min,
+			p->max);
+		goto err_bad_param;
+	}
+	if (p->max > MLXSW_CORE_RES_GET(mlxsw_sp->core, MAX_BUFFER_SIZE)) {
+		dev_err(mlxsw_sp->bus_info->dev,
+			"spectrum: RED: max value %u is too big\n", p->max);
+		goto err_bad_param;
+	}
+	if (p->min == 0 || p->max == 0) {
+		dev_err(mlxsw_sp->bus_info->dev,
+			"spectrum: RED: 0 value is illegal for min and max\n");
+		goto err_bad_param;
+	}
+
+	/* calculate probability in percentage */
+	prob = p->probability;
+	prob *= 100;
+	prob = DIV_ROUND_UP(prob, 1 << 16);
+	prob = DIV_ROUND_UP(prob, 1 << 16);
+	min = mlxsw_sp_bytes_cells(mlxsw_sp, p->min);
+	max = mlxsw_sp_bytes_cells(mlxsw_sp, p->max);
+	err = mlxsw_sp_tclass_congestion_enable(mlxsw_sp_port, tclass_num, min,
+						max, prob, p->is_ecn);
+	if (err)
+		goto err_config;
+
+	mlxsw_sp_qdisc->type = MLXSW_SP_QDISC_RED;
+	if (mlxsw_sp_qdisc->handle != handle)
+		mlxsw_sp_setup_tc_qdisc_clean_stats(mlxsw_sp_port,
+						    mlxsw_sp_qdisc,
+						    tclass_num);
+
+	mlxsw_sp_qdisc->handle = handle;
+	return 0;
+
+err_bad_param:
+	err = -EINVAL;
+err_config:
+	mlxsw_sp_qdisc_red_destroy(mlxsw_sp_port, mlxsw_sp_qdisc->handle,
+				   mlxsw_sp_qdisc, tclass_num);
+	return err;
+}
+
+static int
+mlxsw_sp_qdisc_get_red_xstats(struct mlxsw_sp_port *mlxsw_sp_port, u32 handle,
+			      struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			      int tclass_num, struct red_stats *res)
+{
+	struct red_stats *xstats_base = &mlxsw_sp_qdisc->xstats_base;
+	struct mlxsw_sp_port_xstats *xstats;
+
+	if (mlxsw_sp_qdisc->handle != handle ||
+	    mlxsw_sp_qdisc->type != MLXSW_SP_QDISC_RED)
+		return -EOPNOTSUPP;
+
+	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
+
+	res->prob_drop = xstats->wred_drop[tclass_num] - xstats_base->prob_drop;
+	res->prob_mark = xstats->ecn - xstats_base->prob_mark;
+	res->pdrop = xstats->tail_drop[tclass_num] - xstats_base->pdrop;
+	return 0;
+}
+
+static int
+mlxsw_sp_qdisc_get_red_stats(struct mlxsw_sp_port *mlxsw_sp_port, u32 handle,
+			     struct mlxsw_sp_qdisc *mlxsw_sp_qdisc,
+			     int tclass_num,
+			     struct tc_red_qopt_offload_stats *res)
+{
+	u64 tx_bytes, tx_packets, overlimits, drops;
+	struct mlxsw_sp_port_xstats *xstats;
+	struct rtnl_link_stats64 *stats;
+
+	if (mlxsw_sp_qdisc->handle != handle ||
+	    mlxsw_sp_qdisc->type != MLXSW_SP_QDISC_RED)
+		return -EOPNOTSUPP;
+
+	xstats = &mlxsw_sp_port->periodic_hw_stats.xstats;
+	stats = &mlxsw_sp_port->periodic_hw_stats.stats;
+
+	tx_bytes = stats->tx_bytes - mlxsw_sp_qdisc->tx_bytes;
+	tx_packets = stats->tx_packets - mlxsw_sp_qdisc->tx_packets;
+	overlimits = xstats->wred_drop[tclass_num] + xstats->ecn -
+		     mlxsw_sp_qdisc->overlimits;
+	drops = xstats->wred_drop[tclass_num] + xstats->tail_drop[tclass_num] -
+		mlxsw_sp_qdisc->drops;
+
+	_bstats_update(res->bstats, tx_bytes, tx_packets);
+	res->qstats->overlimits += overlimits;
+	res->qstats->drops += drops;
+	res->qstats->backlog += mlxsw_sp_cells_bytes(mlxsw_sp_port->mlxsw_sp,
+						xstats->backlog[tclass_num]);
+
+	mlxsw_sp_qdisc->drops +=  drops;
+	mlxsw_sp_qdisc->overlimits += overlimits;
+	mlxsw_sp_qdisc->tx_bytes += tx_bytes;
+	mlxsw_sp_qdisc->tx_packets += tx_packets;
+	return 0;
+}
+
+#define MLXSW_SP_PORT_DEFAULT_TCLASS 0
+
+int mlxsw_sp_setup_tc_red(struct mlxsw_sp_port *mlxsw_sp_port,
+			  struct tc_red_qopt_offload *p)
+{
+	struct mlxsw_sp_qdisc *mlxsw_sp_qdisc;
+	int tclass_num;
+
+	if (p->parent != TC_H_ROOT)
+		return -EOPNOTSUPP;
+
+	mlxsw_sp_qdisc = &mlxsw_sp_port->root_qdisc;
+	tclass_num = MLXSW_SP_PORT_DEFAULT_TCLASS;
+
+	switch (p->command) {
+	case TC_RED_REPLACE:
+		return mlxsw_sp_qdisc_red_replace(mlxsw_sp_port, p->handle,
+						  mlxsw_sp_qdisc, tclass_num,
+						  &p->set);
+	case TC_RED_DESTROY:
+		return mlxsw_sp_qdisc_red_destroy(mlxsw_sp_port, p->handle,
+						  mlxsw_sp_qdisc, tclass_num);
+	case TC_RED_XSTATS:
+		return mlxsw_sp_qdisc_get_red_xstats(mlxsw_sp_port, p->handle,
+						     mlxsw_sp_qdisc, tclass_num,
+						     p->xstats);
+	case TC_RED_STATS:
+		return mlxsw_sp_qdisc_get_red_stats(mlxsw_sp_port, p->handle,
+						    mlxsw_sp_qdisc, tclass_num,
+						    &p->stats);
+	default:
+		return -EOPNOTSUPP;
+	}
+}

drivers/net/ethernet/sfc/falcon/tx.c

···
 	unsigned tc, num_tc;
 	int rc;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	num_tc = mqprio->num_tc;

drivers/net/ethernet/sfc/tx.c

···
 	unsigned tc, num_tc;
 	int rc;
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	num_tc = mqprio->num_tc;

drivers/net/ethernet/ti/netcp_core.c

···
 	/* setup tc must be called under rtnl lock */
 	ASSERT_RTNL();
 
-	if (type != TC_SETUP_MQPRIO)
+	if (type != TC_SETUP_QDISC_MQPRIO)
 		return -EOPNOTSUPP;
 
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;

include/linux/netdevice.h

···
 				       struct sk_buff *skb);
 
 enum tc_setup_type {
-	TC_SETUP_MQPRIO,
+	TC_SETUP_QDISC_MQPRIO,
 	TC_SETUP_CLSU32,
 	TC_SETUP_CLSFLOWER,
 	TC_SETUP_CLSMATCHALL,
 	TC_SETUP_CLSBPF,
 	TC_SETUP_BLOCK,
-	TC_SETUP_CBS,
+	TC_SETUP_QDISC_CBS,
+	TC_SETUP_QDISC_RED,
 };
 
 /* These structures hold the attributes of bpf state that are being passed

include/net/pkt_cls.h

···
 	u8  *data;
 	u32 len;
 };
+
+enum tc_red_command {
+	TC_RED_REPLACE,
+	TC_RED_DESTROY,
+	TC_RED_STATS,
+	TC_RED_XSTATS,
+};
+
+struct tc_red_qopt_offload_params {
+	u32 min;
+	u32 max;
+	u32 probability;
+	bool is_ecn;
+};
+struct tc_red_qopt_offload_stats {
+	struct gnet_stats_basic_packed *bstats;
+	struct gnet_stats_queue *qstats;
+};
+
+struct tc_red_qopt_offload {
+	enum tc_red_command command;
+	u32 handle;
+	u32 parent;
+	union {
+		struct tc_red_qopt_offload_params set;
+		struct tc_red_qopt_offload_stats stats;
+		struct red_stats *xstats;
+	};
+};
+
 #endif

include/uapi/linux/pkt_sched.h

···
 #define TC_RED_ECN		1
 #define TC_RED_HARDDROP		2
 #define TC_RED_ADAPTATIVE	4
+#define TC_RED_OFFLOADED	8
 };
 
 struct tc_red_xstats {

net/sched/sch_cbs.c

···
 	cbs.queue = q->queue;
 	cbs.enable = 0;
 
-	err = ops->ndo_setup_tc(dev, TC_SETUP_CBS, &cbs);
+	err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_CBS, &cbs);
 	if (err < 0)
 		pr_warn("Couldn't disable CBS offload for queue %d\n",
 			cbs.queue);
···
 	cbs.idleslope = opt->idleslope;
 	cbs.sendslope = opt->sendslope;
 
-	err = ops->ndo_setup_tc(dev, TC_SETUP_CBS, &cbs);
+	err = ops->ndo_setup_tc(dev, TC_SETUP_QDISC_CBS, &cbs);
 	if (err < 0)
 		return err;
 

net/sched/sch_mqprio.c

···
 		switch (priv->mode) {
 		case TC_MQPRIO_MODE_DCB:
 		case TC_MQPRIO_MODE_CHANNEL:
-			dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_MQPRIO,
+			dev->netdev_ops->ndo_setup_tc(dev,
+						      TC_SETUP_QDISC_MQPRIO,
 						      &mqprio);
 			break;
 		default:
···
 			return -EINVAL;
 		}
 		err = dev->netdev_ops->ndo_setup_tc(dev,
-						    TC_SETUP_MQPRIO,
+						    TC_SETUP_QDISC_MQPRIO,
 						    &mqprio);
 		if (err)
 			return err;

net/sched/sch_red.c

···
 #include <linux/kernel.h>
 #include <linux/skbuff.h>
 #include <net/pkt_sched.h>
+#include <net/pkt_cls.h>
 #include <net/inet_ecn.h>
 #include <net/red.h>
 
···
 	red_restart(&q->vars);
 }
 
+static int red_offload(struct Qdisc *sch, bool enable)
+{
+	struct red_sched_data *q = qdisc_priv(sch);
+	struct net_device *dev = qdisc_dev(sch);
+	struct tc_red_qopt_offload opt = {
+		.handle = sch->handle,
+		.parent = sch->parent,
+	};
+
+	if (!tc_can_offload(dev) || !dev->netdev_ops->ndo_setup_tc)
+		return -EOPNOTSUPP;
+
+	if (enable) {
+		opt.command = TC_RED_REPLACE;
+		opt.set.min = q->parms.qth_min >> q->parms.Wlog;
+		opt.set.max = q->parms.qth_max >> q->parms.Wlog;
+		opt.set.probability = q->parms.max_P;
+		opt.set.is_ecn = red_use_ecn(q);
+	} else {
+		opt.command = TC_RED_DESTROY;
+	}
+
+	return dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_QDISC_RED, &opt);
+}
+
 static void red_destroy(struct Qdisc *sch)
 {
 	struct red_sched_data *q = qdisc_priv(sch);
 
 	del_timer_sync(&q->adapt_timer);
+	red_offload(sch, false);
 	qdisc_destroy(q->qdisc);
 }
 
···
 		red_start_of_idle_period(&q->vars);
 
 	sch_tree_unlock(sch);
+	red_offload(sch, true);
 	return 0;
 }
 
···
 	return red_change(sch, opt);
 }
 
+static int red_dump_offload(struct Qdisc *sch, struct tc_red_qopt *opt)
+{
+	struct net_device *dev = qdisc_dev(sch);
+	struct tc_red_qopt_offload hw_stats = {
+		.handle = sch->handle,
+		.parent = sch->parent,
+		.command = TC_RED_STATS,
+		.stats.bstats = &sch->bstats,
+		.stats.qstats = &sch->qstats,
+	};
+	int err;
+
+	opt->flags &= ~TC_RED_OFFLOADED;
+	if (!tc_can_offload(dev) || !dev->netdev_ops->ndo_setup_tc)
+		return 0;
+
+	err = dev->netdev_ops->ndo_setup_tc(dev, TC_SETUP_QDISC_RED,
+					    &hw_stats);
+	if (err == -EOPNOTSUPP)
+		return 0;
+
+	if (!err)
+		opt->flags |= TC_RED_OFFLOADED;
+
+	return err;
+}
+
 static int red_dump(struct Qdisc *sch, struct sk_buff *skb)
 {
 	struct red_sched_data *q = qdisc_priv(sch);
···
 		.Plog		= q->parms.Plog,
 		.Scell_log	= q->parms.Scell_log,
 	};
+	int err;
 
 	sch->qstats.backlog = q->qdisc->qstats.backlog;
+	err = red_dump_offload(sch, &opt);
+	if (err)
+		goto nla_put_failure;
+
 	opts = nla_nest_start(skb, TCA_OPTIONS);
 	if (opts == NULL)
 		goto nla_put_failure;
···
 static int red_dump_stats(struct Qdisc *sch, struct gnet_dump *d)
 {
 	struct red_sched_data *q = qdisc_priv(sch);
+	struct net_device *dev = qdisc_dev(sch);
 	struct tc_red_xstats st = {
 		.early	= q->stats.prob_drop + q->stats.forced_drop,
 		.pdrop	= q->stats.pdrop,
 		.other	= q->stats.other,
 		.marked	= q->stats.prob_mark + q->stats.forced_mark,
 	};
+
+	if (tc_can_offload(dev) &&  dev->netdev_ops->ndo_setup_tc) {
+		struct red_stats hw_stats = {0};
+		struct tc_red_qopt_offload hw_stats_request = {
+			.handle = sch->handle,
+			.parent = sch->parent,
+			.command = TC_RED_XSTATS,
+			.xstats = &hw_stats,
+		};
+		if (!dev->netdev_ops->ndo_setup_tc(dev,
+						   TC_SETUP_QDISC_RED,
+						   &hw_stats_request)) {
+			st.early += hw_stats.prob_drop + hw_stats.forced_drop;
+			st.pdrop += hw_stats.pdrop;
+			st.other += hw_stats.other;
+			st.marked += hw_stats.prob_mark + hw_stats.forced_mark;
+		}
+	}
 
 	return gnet_stats_copy_app(d, &st, sizeof(st));
 }