[PATCH v3] net/mlx5: support multiple groups and jump action

[Yongseok Koh <yskoh@mellanox.com>]

rte_flow has 'group' attribute and 'jump' action in order to support
multiple groups. This feature is known as multi-table support ('chain' in
linux TC flower) in general because a group means a table of flows. Example
commands are:

	flow create 0 transfer priority 1 ingress
	     pattern eth / vlan vid is 100 / end
	     actions jump group 1 / end

	flow create 0 transfer priority 1 ingress
	     pattern eth / vlan vid is 200 / end
	     actions jump group 2 / end

	flow create 0 transfer group 1 priority 2 ingress
	     pattern eth / vlan vid is 100 /
	     	     ipv4 dst spec 192.168.40.0 dst prefix 24 / end
	     actions drop / end

	flow create 0 transfer group 1 priority 2 ingress
	     pattern end
	     actions of_pop_vlan / port_id id 1 / end

	flow create 0 transfer group 2 priority 2 ingress
	     pattern eth / vlan vid is 200 /
	     	     ipv4 dst spec 192.168.40.0 dst prefix 24 / end
	     actions of_pop_vlan / port_id id 2 / end

	flow create 0 transfer group 2 priority 2 ingress
	     pattern end
	     actions port_id id 2 / end

With theses flows, if a packet having vlan 200 and src_ip as 192.168.40.1,
this packet will firstly hit the 1st flow. Then it will hit the 5th flow
because of the 'jump' action. As a result, the packet will be forwarded to
port 2 (VF representor) with vlan tag being stripped off. If the packet had
vlan 100 instead, it would be dropped by the 3rd flow.

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
Acked-by: Ori Kam <orika@mellanox.com>
---

v3:
* reorder build macros

v2:
* drop ethdev patch as it had already been fixed by Adrien's patch
* move TCF macros from mlx5_flow.h to mlx5_flow_tcf.c

 drivers/net/mlx5/Makefile        | 10 +++++
 drivers/net/mlx5/meson.build     |  4 ++
 drivers/net/mlx5/mlx5_flow.h     |  1 +
 drivers/net/mlx5/mlx5_flow_tcf.c | 82 +++++++++++++++++++++++++++++++++++-----
 4 files changed, 88 insertions(+), 9 deletions(-)

diff --git a/drivers/net/mlx5/Makefile b/drivers/net/mlx5/Makefile
index ca1de9f210..928837bb85 100644
--- a/drivers/net/mlx5/Makefile
+++ b/drivers/net/mlx5/Makefile
@@ -207,6 +207,11 @@ mlx5_autoconf.h.new: $(RTE_SDK)/buildtools/auto-config-h.sh
 		enum IFLA_PHYS_PORT_NAME \
 		$(AUTOCONF_OUTPUT)
 	$Q sh -- '$<' '$@' \
+		HAVE_TCA_CHAIN \
+		linux/rtnetlink.h \
+		enum TCA_CHAIN \
+		$(AUTOCONF_OUTPUT)
+	$Q sh -- '$<' '$@' \
 		HAVE_TCA_FLOWER_ACT \
 		linux/pkt_cls.h \
 		enum TCA_FLOWER_ACT \
@@ -342,6 +347,11 @@ mlx5_autoconf.h.new: $(RTE_SDK)/buildtools/auto-config-h.sh
 		enum TCA_FLOWER_KEY_VLAN_ETH_TYPE \
 		$(AUTOCONF_OUTPUT)
 	$Q sh -- '$<' '$@' \
+		HAVE_TC_ACT_GOTO_CHAIN \
+		linux/pkt_cls.h \
+		define TC_ACT_GOTO_CHAIN \
+		$(AUTOCONF_OUTPUT)
+	$Q sh -- '$<' '$@' \
 		HAVE_TC_ACT_VLAN \
 		linux/tc_act/tc_vlan.h \
 		enum TCA_VLAN_PUSH_VLAN_PRIORITY \
diff --git a/drivers/net/mlx5/meson.build b/drivers/net/mlx5/meson.build
index fd93ac1625..025dcc9e96 100644
--- a/drivers/net/mlx5/meson.build
+++ b/drivers/net/mlx5/meson.build
@@ -126,6 +126,8 @@ if build
 		'IFLA_PHYS_SWITCH_ID' ],
 		[ 'HAVE_IFLA_PHYS_PORT_NAME', 'linux/if_link.h',
 		'IFLA_PHYS_PORT_NAME' ],
+		[ 'HAVE_TCA_CHAIN', 'linux/rtnetlink.h',
+		'TCA_CHAIN' ],
 		[ 'HAVE_TCA_FLOWER_ACT', 'linux/pkt_cls.h',
 		'TCA_FLOWER_ACT' ],
 		[ 'HAVE_TCA_FLOWER_FLAGS', 'linux/pkt_cls.h',
@@ -180,6 +182,8 @@ if build
 		'TCA_FLOWER_KEY_VLAN_PRIO' ],
 		[ 'HAVE_TCA_FLOWER_KEY_VLAN_ETH_TYPE', 'linux/pkt_cls.h',
 		'TCA_FLOWER_KEY_VLAN_ETH_TYPE' ],
+		[ 'HAVE_TC_ACT_GOTO_CHAIN', 'linux/pkt_cls.h',
+		'TC_ACT_GOTO_CHAIN' ],
 		[ 'HAVE_TC_ACT_VLAN', 'linux/tc_act/tc_vlan.h',
 		'TCA_VLAN_PUSH_VLAN_PRIORITY' ],
 		[ 'HAVE_RDMA_NL_NLDEV', 'rdma/rdma_netlink.h',
diff --git a/drivers/net/mlx5/mlx5_flow.h b/drivers/net/mlx5/mlx5_flow.h
index 12de841e86..3ed0ddc585 100644
--- a/drivers/net/mlx5/mlx5_flow.h
+++ b/drivers/net/mlx5/mlx5_flow.h
@@ -78,6 +78,7 @@
 #define MLX5_FLOW_ACTION_OF_PUSH_VLAN (1u << 8)
 #define MLX5_FLOW_ACTION_OF_SET_VLAN_VID (1u << 9)
 #define MLX5_FLOW_ACTION_OF_SET_VLAN_PCP (1u << 10)
+#define MLX5_FLOW_ACTION_JUMP (1u << 11)
 
 #define MLX5_FLOW_FATE_ACTIONS \
 	(MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_QUEUE | MLX5_FLOW_ACTION_RSS)
diff --git a/drivers/net/mlx5/mlx5_flow_tcf.c b/drivers/net/mlx5/mlx5_flow_tcf.c
index 91f6ef6786..fbc4c2bb7c 100644
--- a/drivers/net/mlx5/mlx5_flow_tcf.c
+++ b/drivers/net/mlx5/mlx5_flow_tcf.c
@@ -148,6 +148,12 @@ struct tc_vlan {
 #ifndef HAVE_TCA_FLOWER_KEY_VLAN_ETH_TYPE
 #define TCA_FLOWER_KEY_VLAN_ETH_TYPE 25
 #endif
+#ifndef HAVE_TCA_CHAIN
+#define TCA_CHAIN 11
+#endif
+#ifndef HAVE_TC_ACT_GOTO_CHAIN
+#define TC_ACT_GOTO_CHAIN 0x20000000
+#endif
 
 #ifndef IPV6_ADDR_LEN
 #define IPV6_ADDR_LEN 16
@@ -225,7 +231,13 @@ struct flow_tcf_ptoi {
 	unsigned int ifindex; /**< Network interface index. */
 };
 
-#define MLX5_TCF_FATE_ACTIONS (MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_PORT_ID)
+/* Due to a limitation on driver/FW. */
+#define MLX5_TCF_GROUP_ID_MAX 3
+#define MLX5_TCF_GROUP_PRIORITY_MAX 14
+
+#define MLX5_TCF_FATE_ACTIONS \
+	(MLX5_FLOW_ACTION_DROP | MLX5_FLOW_ACTION_PORT_ID | \
+	 MLX5_FLOW_ACTION_JUMP)
 #define MLX5_TCF_VLAN_ACTIONS \
 	(MLX5_FLOW_ACTION_OF_POP_VLAN | MLX5_FLOW_ACTION_OF_PUSH_VLAN | \
 	 MLX5_FLOW_ACTION_OF_SET_VLAN_VID | MLX5_FLOW_ACTION_OF_SET_VLAN_PCP)
@@ -370,14 +382,25 @@ flow_tcf_validate_attributes(const struct rte_flow_attr *attr,
 			     struct rte_flow_error *error)
 {
 	/*
-	 * Supported attributes: no groups, some priorities and ingress only.
-	 * Don't care about transfer as it is the caller's problem.
+	 * Supported attributes: groups, some priorities and ingress only.
+	 * group is supported only if kernel supports chain. Don't care about
+	 * transfer as it is the caller's problem.
 	 */
-	if (attr->group)
+	if (attr->group > MLX5_TCF_GROUP_ID_MAX)
 		return rte_flow_error_set(error, ENOTSUP,
 					  RTE_FLOW_ERROR_TYPE_ATTR_GROUP, attr,
-					  "groups are not supported");
-	if (attr->priority > 0xfffe)
+					  "group ID larger than "
+					  RTE_STR(MLX5_TCF_GROUP_ID_MAX)
+					  " isn't supported");
+	else if (attr->group > 0 &&
+		 attr->priority > MLX5_TCF_GROUP_PRIORITY_MAX)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY,
+					  attr,
+					  "lowest priority level is "
+					  RTE_STR(MLX5_TCF_GROUP_PRIORITY_MAX)
+					  " when group is configured");
+	else if (attr->priority > 0xfffe)
 		return rte_flow_error_set(error, ENOTSUP,
 					  RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY,
 					  attr,
@@ -428,6 +451,7 @@ flow_tcf_validate(struct rte_eth_dev *dev,
 	} spec, mask;
 	union {
 		const struct rte_flow_action_port_id *port_id;
+		const struct rte_flow_action_jump *jump;
 		const struct rte_flow_action_of_push_vlan *of_push_vlan;
 		const struct rte_flow_action_of_set_vlan_vid *
 			of_set_vlan_vid;
@@ -675,6 +699,16 @@ flow_tcf_validate(struct rte_eth_dev *dev,
 			action_flags |= MLX5_FLOW_ACTION_PORT_ID;
 			port_id_dev = &rte_eth_devices[conf.port_id->id];
 			break;
+		case RTE_FLOW_ACTION_TYPE_JUMP:
+			conf.jump = actions->conf;
+			if (attr->group >= conf.jump->group)
+				return rte_flow_error_set
+					(error, ENOTSUP,
+					 RTE_FLOW_ERROR_TYPE_ACTION,
+					 actions,
+					 "can't jump to a group backward");
+			action_flags |= MLX5_FLOW_ACTION_JUMP;
+			break;
 		case RTE_FLOW_ACTION_TYPE_DROP:
 			if (action_flags & MLX5_TCF_FATE_ACTIONS)
 				return rte_flow_error_set
@@ -757,7 +791,8 @@ flow_tcf_validate(struct rte_eth_dev *dev,
  *   Maximum size of memory for items.
  */
 static int
-flow_tcf_get_items_and_size(const struct rte_flow_item items[],
+flow_tcf_get_items_and_size(const struct rte_flow_attr *attr,
+			    const struct rte_flow_item items[],
 			    uint64_t *item_flags)
 {
 	int size = 0;
@@ -766,6 +801,8 @@ flow_tcf_get_items_and_size(const struct rte_flow_item items[],
 	size += SZ_NLATTR_STRZ_OF("flower") +
 		SZ_NLATTR_NEST + /* TCA_OPTIONS. */
 		SZ_NLATTR_TYPE_OF(uint32_t); /* TCA_CLS_FLAGS_SKIP_SW. */
+	if (attr->group > 0)
+		size += SZ_NLATTR_TYPE_OF(uint32_t); /* TCA_CHAIN. */
 	for (; items->type != RTE_FLOW_ITEM_TYPE_END; items++) {
 		switch (items->type) {
 		case RTE_FLOW_ITEM_TYPE_VOID:
@@ -855,6 +892,13 @@ flow_tcf_get_actions_and_size(const struct rte_flow_action actions[],
 				SZ_NLATTR_TYPE_OF(struct tc_mirred);
 			flags |= MLX5_FLOW_ACTION_PORT_ID;
 			break;
+		case RTE_FLOW_ACTION_TYPE_JUMP:
+			size += SZ_NLATTR_NEST + /* na_act_index. */
+				SZ_NLATTR_STRZ_OF("gact") +
+				SZ_NLATTR_NEST + /* TCA_ACT_OPTIONS. */
+				SZ_NLATTR_TYPE_OF(struct tc_gact);
+			flags |= MLX5_FLOW_ACTION_JUMP;
+			break;
 		case RTE_FLOW_ACTION_TYPE_DROP:
 			size += SZ_NLATTR_NEST + /* na_act_index. */
 				SZ_NLATTR_STRZ_OF("gact") +
@@ -940,7 +984,7 @@ flow_tcf_nl_brand(struct nlmsghdr *nlh, uint32_t handle)
  *   otherwise NULL and rte_ernno is set.
  */
 static struct mlx5_flow *
-flow_tcf_prepare(const struct rte_flow_attr *attr __rte_unused,
+flow_tcf_prepare(const struct rte_flow_attr *attr,
 		 const struct rte_flow_item items[],
 		 const struct rte_flow_action actions[],
 		 uint64_t *item_flags, uint64_t *action_flags,
@@ -953,7 +997,7 @@ flow_tcf_prepare(const struct rte_flow_attr *attr __rte_unused,
 	struct nlmsghdr *nlh;
 	struct tcmsg *tcm;
 
-	size += flow_tcf_get_items_and_size(items, item_flags);
+	size += flow_tcf_get_items_and_size(attr, items, item_flags);
 	size += flow_tcf_get_actions_and_size(actions, action_flags);
 	dev_flow = rte_zmalloc(__func__, size, MNL_ALIGNTO);
 	if (!dev_flow) {
@@ -1024,6 +1068,7 @@ flow_tcf_translate(struct rte_eth_dev *dev, struct mlx5_flow *dev_flow,
 	} spec, mask;
 	union {
 		const struct rte_flow_action_port_id *port_id;
+		const struct rte_flow_action_jump *jump;
 		const struct rte_flow_action_of_push_vlan *of_push_vlan;
 		const struct rte_flow_action_of_set_vlan_vid *
 			of_set_vlan_vid;
@@ -1058,6 +1103,8 @@ flow_tcf_translate(struct rte_eth_dev *dev, struct mlx5_flow *dev_flow,
 	 */
 	tcm->tcm_info = TC_H_MAKE((attr->priority + 1) << 16,
 				  RTE_BE16(ETH_P_ALL));
+	if (attr->group > 0)
+		mnl_attr_put_u32(nlh, TCA_CHAIN, attr->group);
 	mnl_attr_put_strz(nlh, TCA_KIND, "flower");
 	na_flower = mnl_attr_nest_start(nlh, TCA_OPTIONS);
 	mnl_attr_put_u32(nlh, TCA_FLOWER_FLAGS, TCA_CLS_FLAGS_SKIP_SW);
@@ -1332,6 +1379,23 @@ flow_tcf_translate(struct rte_eth_dev *dev, struct mlx5_flow *dev_flow,
 			mnl_attr_nest_end(nlh, na_act);
 			mnl_attr_nest_end(nlh, na_act_index);
 			break;
+		case RTE_FLOW_ACTION_TYPE_JUMP:
+			conf.jump = actions->conf;
+			na_act_index =
+				mnl_attr_nest_start(nlh, na_act_index_cur++);
+			assert(na_act_index);
+			mnl_attr_put_strz(nlh, TCA_ACT_KIND, "gact");
+			na_act = mnl_attr_nest_start(nlh, TCA_ACT_OPTIONS);
+			assert(na_act);
+			mnl_attr_put(nlh, TCA_GACT_PARMS,
+				     sizeof(struct tc_gact),
+				     &(struct tc_gact){
+					.action = TC_ACT_GOTO_CHAIN |
+						  conf.jump->group,
+				     });
+			mnl_attr_nest_end(nlh, na_act);
+			mnl_attr_nest_end(nlh, na_act_index);
+			break;
 		case RTE_FLOW_ACTION_TYPE_DROP:
 			na_act_index =
 				mnl_attr_nest_start(nlh, na_act_index_cur++);
-- 
2.11.0