net/mlx5e: Create aRFS flow tables

Create the following four flow tables for aRFS usage:
1. IPv4 TCP - filtering 4-tuple of IPv4 TCP packets.
2. IPv6 TCP - filtering 4-tuple of IPv6 TCP packets.
3. IPv4 UDP - filtering 4-tuple of IPv4 UDP packets.
4. IPv6 UDP - filtering 4-tuple of IPv6 UDP packets.

Each flow table has two flow groups: one for the 4-tuple
filtering (full match)  and the other contains * rule for miss rule.

Full match rule means a hit for aRFS and packet will be forwarded
to the dedicated RQ/Core, miss rule packets will be forwarded to
default RSS hashing.

Signed-off-by: Maor Gottlieb <maorg@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 313 insertions, 14 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/Makefile b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
index 4fc45ee0c5d1..679e18ffb3a6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/Makefile
+++ b/drivers/net/ethernet/mellanox/mlx5/core/Makefile
@@ -9,3 +9,4 @@ mlx5_core-$(CONFIG_MLX5_CORE_EN) += wq.o eswitch.o \
 		en_txrx.o en_clock.o vxlan.o en_tc.o
 
 mlx5_core-$(CONFIG_MLX5_CORE_EN_DCB) +=  en_dcbnl.o
+mlx5_core-$(CONFIG_RFS_ACCEL) +=  en_arfs.o
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
index 2c9879c011b1..999e05826490 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -48,6 +48,8 @@
 #include "mlx5_core.h"
 #include "en_stats.h"
 
+#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
+
 #define MLX5E_MAX_NUM_TC	8
 
 #define MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE                0x6
@@ -446,12 +448,38 @@ struct mlx5e_ttc_table {
 	struct mlx5_flow_rule	 *rules[MLX5E_NUM_TT];
 };
 
+struct arfs_table {
+	struct mlx5e_flow_table  ft;
+	struct mlx5_flow_rule    *default_rule;
+};
+
+enum  arfs_type {
+	ARFS_IPV4_TCP,
+	ARFS_IPV6_TCP,
+	ARFS_IPV4_UDP,
+	ARFS_IPV6_UDP,
+	ARFS_NUM_TYPES,
+};
+
+struct mlx5e_arfs_tables {
+	struct arfs_table arfs_tables[ARFS_NUM_TYPES];
+};
+
+/* NIC prio FTS */
+enum {
+	MLX5E_VLAN_FT_LEVEL = 0,
+	MLX5E_L2_FT_LEVEL,
+	MLX5E_TTC_FT_LEVEL,
+	MLX5E_ARFS_FT_LEVEL
+};
+
 struct mlx5e_flow_steering {
 	struct mlx5_flow_namespace      *ns;
 	struct mlx5e_tc_table           tc;
 	struct mlx5e_vlan_table         vlan;
 	struct mlx5e_l2_table           l2;
 	struct mlx5e_ttc_table          ttc;
+	struct mlx5e_arfs_tables        arfs;
 };
 
 struct mlx5e_direct_tir {
@@ -570,6 +598,7 @@ void mlx5e_update_stats(struct mlx5e_priv *priv);
 int mlx5e_create_flow_steering(struct mlx5e_priv *priv);
 void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv);
 void mlx5e_init_l2_addr(struct mlx5e_priv *priv);
+void mlx5e_destroy_flow_table(struct mlx5e_flow_table *ft);
 void mlx5e_set_rx_mode_work(struct work_struct *work);
 
 void mlx5e_fill_hwstamp(struct mlx5e_tstamp *clock, u64 timestamp,
@@ -646,6 +675,18 @@ extern const struct dcbnl_rtnl_ops mlx5e_dcbnl_ops;
 int mlx5e_dcbnl_ieee_setets_core(struct mlx5e_priv *priv, struct ieee_ets *ets);
 #endif
 
+#ifndef CONFIG_RFS_ACCEL
+static inline int mlx5e_arfs_create_tables(struct mlx5e_priv *priv)
+{
+	return 0;
+}
+
+static inline void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv) {}
+#else
+int mlx5e_arfs_create_tables(struct mlx5e_priv *priv);
+void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv);
+#endif
+
 u16 mlx5e_get_max_inline_cap(struct mlx5_core_dev *mdev);
 
 #endif /* __MLX5_EN_H__ */
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c
new file mode 100644
index 000000000000..cd504197855b
--- /dev/null
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_arfs.c
@@ -0,0 +1,251 @@
+/*
+ * Copyright (c) 2016, Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - 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.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include "en.h"
+#include <linux/mlx5/fs.h>
+
+static void arfs_destroy_table(struct arfs_table *arfs_t)
+{
+	mlx5_del_flow_rule(arfs_t->default_rule);
+	mlx5e_destroy_flow_table(&arfs_t->ft);
+}
+
+void mlx5e_arfs_destroy_tables(struct mlx5e_priv *priv)
+{
+	int i;
+
+	if (!(priv->netdev->hw_features & NETIF_F_NTUPLE))
+		return;
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		if (!IS_ERR_OR_NULL(priv->fs.arfs.arfs_tables[i].ft.t))
+			arfs_destroy_table(&priv->fs.arfs.arfs_tables[i]);
+	}
+}
+
+static int arfs_add_default_rule(struct mlx5e_priv *priv,
+				 enum arfs_type type)
+{
+	struct arfs_table *arfs_t = &priv->fs.arfs.arfs_tables[type];
+	struct mlx5_flow_destination dest;
+	u8 match_criteria_enable = 0;
+	u32 *tirn = priv->indir_tirn;
+	u32 *match_criteria;
+	u32 *match_value;
+	int err = 0;
+
+	match_value	= mlx5_vzalloc(MLX5_ST_SZ_BYTES(fte_match_param));
+	match_criteria	= mlx5_vzalloc(MLX5_ST_SZ_BYTES(fte_match_param));
+	if (!match_value || !match_criteria) {
+		netdev_err(priv->netdev, "%s: alloc failed\n", __func__);
+		err = -ENOMEM;
+		goto out;
+	}
+
+	dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;
+	switch (type) {
+	case ARFS_IPV4_TCP:
+		dest.tir_num = tirn[MLX5E_TT_IPV4_TCP];
+		break;
+	case ARFS_IPV4_UDP:
+		dest.tir_num = tirn[MLX5E_TT_IPV4_UDP];
+		break;
+	case ARFS_IPV6_TCP:
+		dest.tir_num = tirn[MLX5E_TT_IPV6_TCP];
+		break;
+	case ARFS_IPV6_UDP:
+		dest.tir_num = tirn[MLX5E_TT_IPV6_UDP];
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	arfs_t->default_rule = mlx5_add_flow_rule(arfs_t->ft.t, match_criteria_enable,
+						  match_criteria, match_value,
+						  MLX5_FLOW_CONTEXT_ACTION_FWD_DEST,
+						  MLX5_FS_DEFAULT_FLOW_TAG,
+						  &dest);
+	if (IS_ERR(arfs_t->default_rule)) {
+		err = PTR_ERR(arfs_t->default_rule);
+		arfs_t->default_rule = NULL;
+		netdev_err(priv->netdev, "%s: add rule failed, arfs type=%d\n",
+			   __func__, type);
+	}
+out:
+	kvfree(match_criteria);
+	kvfree(match_value);
+	return err;
+}
+
+#define MLX5E_ARFS_NUM_GROUPS	2
+#define MLX5E_ARFS_GROUP1_SIZE	BIT(12)
+#define MLX5E_ARFS_GROUP2_SIZE	BIT(0)
+#define MLX5E_ARFS_TABLE_SIZE	(MLX5E_ARFS_GROUP1_SIZE +\
+				 MLX5E_ARFS_GROUP2_SIZE)
+static int arfs_create_groups(struct mlx5e_flow_table *ft,
+			      enum  arfs_type type)
+{
+	int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
+	void *outer_headers_c;
+	int ix = 0;
+	u32 *in;
+	int err;
+	u8 *mc;
+
+	ft->g = kcalloc(MLX5E_ARFS_NUM_GROUPS,
+			sizeof(*ft->g), GFP_KERNEL);
+	in = mlx5_vzalloc(inlen);
+	if  (!in || !ft->g) {
+		kvfree(ft->g);
+		kvfree(in);
+		return -ENOMEM;
+	}
+
+	mc = MLX5_ADDR_OF(create_flow_group_in, in, match_criteria);
+	outer_headers_c = MLX5_ADDR_OF(fte_match_param, mc,
+				       outer_headers);
+	MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, ethertype);
+	switch (type) {
+	case ARFS_IPV4_TCP:
+	case ARFS_IPV6_TCP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, tcp_dport);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, tcp_sport);
+		break;
+	case ARFS_IPV4_UDP:
+	case ARFS_IPV6_UDP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, udp_dport);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c, udp_sport);
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	switch (type) {
+	case ARFS_IPV4_TCP:
+	case ARFS_IPV4_UDP:
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c,
+				 src_ipv4_src_ipv6.ipv4_layout.ipv4);
+		MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, outer_headers_c,
+				 dst_ipv4_dst_ipv6.ipv4_layout.ipv4);
+		break;
+	case ARFS_IPV6_TCP:
+	case ARFS_IPV6_UDP:
+		memset(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
+				    src_ipv4_src_ipv6.ipv6_layout.ipv6),
+		       0xff, 16);
+		memset(MLX5_ADDR_OF(fte_match_set_lyr_2_4, outer_headers_c,
+				    dst_ipv4_dst_ipv6.ipv6_layout.ipv6),
+		       0xff, 16);
+		break;
+	default:
+		err = -EINVAL;
+		goto out;
+	}
+
+	MLX5_SET_CFG(in, match_criteria_enable, MLX5_MATCH_OUTER_HEADERS);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_ARFS_GROUP1_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	memset(in, 0, inlen);
+	MLX5_SET_CFG(in, start_flow_index, ix);
+	ix += MLX5E_ARFS_GROUP2_SIZE;
+	MLX5_SET_CFG(in, end_flow_index, ix - 1);
+	ft->g[ft->num_groups] = mlx5_create_flow_group(ft->t, in);
+	if (IS_ERR(ft->g[ft->num_groups]))
+		goto err;
+	ft->num_groups++;
+
+	kvfree(in);
+	return 0;
+
+err:
+	err = PTR_ERR(ft->g[ft->num_groups]);
+	ft->g[ft->num_groups] = NULL;
+out:
+	kvfree(in);
+
+	return err;
+}
+
+static int arfs_create_table(struct mlx5e_priv *priv,
+			     enum arfs_type type)
+{
+	struct mlx5e_arfs_tables *arfs = &priv->fs.arfs;
+	struct mlx5e_flow_table *ft = &arfs->arfs_tables[type].ft;
+	int err;
+
+	ft->t = mlx5_create_flow_table(priv->fs.ns, MLX5E_NIC_PRIO,
+				       MLX5E_ARFS_TABLE_SIZE, MLX5E_ARFS_FT_LEVEL);
+	if (IS_ERR(ft->t)) {
+		err = PTR_ERR(ft->t);
+		ft->t = NULL;
+		return err;
+	}
+
+	err = arfs_create_groups(ft, type);
+	if (err)
+		goto err;
+
+	err = arfs_add_default_rule(priv, type);
+	if (err)
+		goto err;
+
+	return 0;
+err:
+	mlx5e_destroy_flow_table(ft);
+	return err;
+}
+
+int mlx5e_arfs_create_tables(struct mlx5e_priv *priv)
+{
+	int err = 0;
+	int i;
+
+	if (!(priv->netdev->hw_features & NETIF_F_NTUPLE))
+		return 0;
+
+	for (i = 0; i < ARFS_NUM_TYPES; i++) {
+		err = arfs_create_table(priv, i);
+		if (err)
+			goto err;
+	}
+	return 0;
+err:
+	mlx5e_arfs_destroy_tables(priv);
+	return err;
+}
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
index 6e353b3a1422..b32740092854 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_fs.c
@@ -42,15 +42,6 @@ static int mlx5e_add_l2_flow_rule(struct mlx5e_priv *priv,
 static void mlx5e_del_l2_flow_rule(struct mlx5e_priv *priv,
 				   struct mlx5e_l2_rule *ai);
 
-/* NIC prio FTS */
-enum {
-	MLX5E_VLAN_FT_LEVEL = 0,
-	MLX5E_L2_FT_LEVEL,
-	MLX5E_TTC_FT_LEVEL
-};
-
-#define MLX5_SET_CFG(p, f, v) MLX5_SET(create_flow_group_in, p, f, v)
-
 enum {
 	MLX5E_FULLMATCH = 0,
 	MLX5E_ALLMULTI  = 1,
@@ -530,7 +521,7 @@ void mlx5e_init_l2_addr(struct mlx5e_priv *priv)
 	ether_addr_copy(priv->fs.l2.broadcast.addr, priv->netdev->broadcast);
 }
 
-static void mlx5e_destroy_flow_table(struct mlx5e_flow_table *ft)
+void mlx5e_destroy_flow_table(struct mlx5e_flow_table *ft)
 {
 	mlx5e_destroy_groups(ft);
 	kfree(ft->g);
@@ -1083,11 +1074,18 @@ int mlx5e_create_flow_steering(struct mlx5e_priv *priv)
 	if (!priv->fs.ns)
 		return -EINVAL;
 
+	err = mlx5e_arfs_create_tables(priv);
+	if (err) {
+		netdev_err(priv->netdev, "Failed to create arfs tables, err=%d\n",
+			   err);
+		priv->netdev->hw_features &= ~NETIF_F_NTUPLE;
+	}
+
 	err = mlx5e_create_ttc_table(priv);
 	if (err) {
 		netdev_err(priv->netdev, "Failed to create ttc table, err=%d\n",
 			   err);
-		return err;
+		goto err_destroy_arfs_tables;
 	}
 
 	err = mlx5e_create_l2_table(priv);
@@ -1110,6 +1108,8 @@ err_destroy_l2_table:
 	mlx5e_destroy_l2_table(priv);
 err_destroy_ttc_table:
 	mlx5e_destroy_ttc_table(priv);
+err_destroy_arfs_tables:
+	mlx5e_arfs_destroy_tables(priv);
 
 	return err;
 }
@@ -1120,4 +1120,5 @@ void mlx5e_destroy_flow_steering(struct mlx5e_priv *priv)
 	mlx5e_destroy_vlan_table(priv);
 	mlx5e_destroy_l2_table(priv);
 	mlx5e_destroy_ttc_table(priv);
+	mlx5e_arfs_destroy_tables(priv);
 }
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
index 8fee224fb98c..20167b9403b6 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c
@@ -2803,8 +2803,12 @@ static void mlx5e_build_netdev(struct net_device *netdev)
 	if (FT_CAP(flow_modify_en) &&
 	    FT_CAP(modify_root) &&
 	    FT_CAP(identified_miss_table_mode) &&
-	    FT_CAP(flow_table_modify))
-		priv->netdev->hw_features      |= NETIF_F_HW_TC;
+	    FT_CAP(flow_table_modify)) {
+		netdev->hw_features      |= NETIF_F_HW_TC;
+#ifdef CONFIG_RFS_ACCEL
+		netdev->hw_features	 |= NETIF_F_NTUPLE;
+#endif
+	}
 
 	netdev->features         |= NETIF_F_HIGHDMA;
 
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
index 83fe8643d89a..4d78d5a48af3 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c
@@ -74,7 +74,8 @@
 #define BY_PASS_MIN_LEVEL (KERNEL_MIN_LEVEL + MLX5_BY_PASS_NUM_PRIOS +\
 			   LEFTOVERS_NUM_PRIOS)
 
-#define KERNEL_NIC_PRIO_NUM_LEVELS 3
+/* Vlan, mac, ttc, aRFS */
+#define KERNEL_NIC_PRIO_NUM_LEVELS 4
 #define KERNEL_NIC_NUM_PRIOS 1
 /* One more level for tc */
 #define KERNEL_MIN_LEVEL (KERNEL_NIC_PRIO_NUM_LEVELS + 1)