[PATCH v2 16/20] net/mlx5: add flow VXLAN item

[Nelio Laranjeiro <nelio.laranjeiro@6wind.com>]

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx5/mlx5_flow.c | 242 +++++++++++++++++++++++++++++------
 1 file changed, 202 insertions(+), 40 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index eedf0c461..daf5b9b5a 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -35,18 +35,45 @@
 extern const struct eth_dev_ops mlx5_dev_ops;
 extern const struct eth_dev_ops mlx5_dev_ops_isolate;
 
-/* Pattern Layer bits. */
+/* Pattern outer Layer bits. */
 #define MLX5_FLOW_LAYER_OUTER_L2 (1u << 0)
 #define MLX5_FLOW_LAYER_OUTER_L3_IPV4 (1u << 1)
 #define MLX5_FLOW_LAYER_OUTER_L3_IPV6 (1u << 2)
 #define MLX5_FLOW_LAYER_OUTER_L4_UDP (1u << 3)
 #define MLX5_FLOW_LAYER_OUTER_L4_TCP (1u << 4)
 #define MLX5_FLOW_LAYER_OUTER_VLAN (1u << 5)
-/* Masks. */
+
+/* Pattern inner Layer bits. */
+#define MLX5_FLOW_LAYER_INNER_L2 (1u << 6)
+#define MLX5_FLOW_LAYER_INNER_L3_IPV4 (1u << 7)
+#define MLX5_FLOW_LAYER_INNER_L3_IPV6 (1u << 8)
+#define MLX5_FLOW_LAYER_INNER_L4_UDP (1u << 9)
+#define MLX5_FLOW_LAYER_INNER_L4_TCP (1u << 10)
+#define MLX5_FLOW_LAYER_INNER_VLAN (1u << 11)
+
+/* Pattern tunnel Layer bits. */
+#define MLX5_FLOW_LAYER_VXLAN (1u << 12)
+
+/* Outer Masks. */
 #define MLX5_FLOW_LAYER_OUTER_L3 \
 	(MLX5_FLOW_LAYER_OUTER_L3_IPV4 | MLX5_FLOW_LAYER_OUTER_L3_IPV6)
 #define MLX5_FLOW_LAYER_OUTER_L4 \
 	(MLX5_FLOW_LAYER_OUTER_L4_UDP | MLX5_FLOW_LAYER_OUTER_L4_TCP)
+#define MLX5_FLOW_LAYER_OUTER \
+	(MLX5_FLOW_LAYER_OUTER_L2 | MLX5_FLOW_LAYER_OUTER_L3 | \
+	 MLX5_FLOW_LAYER_OUTER_L4)
+
+/* Tunnel masks. */
+#define MLX5_FLOW_LAYER_TUNNEL MLX5_FLOW_LAYER_VXLAN
+
+/* Inner Masks. */
+#define MLX5_FLOW_LAYER_INNER_L3 \
+	(MLX5_FLOW_LAYER_INNER_L3_IPV4 | MLX5_FLOW_LAYER_INNER_L3_IPV6)
+#define MLX5_FLOW_LAYER_INNER_L4 \
+	(MLX5_FLOW_LAYER_INNER_L4_UDP | MLX5_FLOW_LAYER_INNER_L4_TCP)
+#define MLX5_FLOW_LAYER_INNER \
+	(MLX5_FLOW_LAYER_INNER_L2 | MLX5_FLOW_LAYER_INNER_L3 | \
+	 MLX5_FLOW_LAYER_INNER_L4)
 
 /* Action fate on the packet. */
 #define MLX5_FLOW_FATE_DROP (1u << 0)
@@ -406,10 +433,14 @@ mlx5_flow_print(struct rte_flow *flow __rte_unused)
 	LIST_FOREACH(verbs, &flow->verbs, next) {
 		uint32_t layers = flow->layers | verbs->layers;
 
-		fprintf(stdout, " layers: %s/%s/%s\n",
+		fprintf(stdout, " layers: %s/%s/%s/%s/%s/%s/%s\n",
 			layers & MLX5_FLOW_LAYER_OUTER_L2 ? "l2" : "-",
 			layers & MLX5_FLOW_LAYER_OUTER_L3 ? "l3" : "-",
-			layers & MLX5_FLOW_LAYER_OUTER_L4 ? "l4" : "-");
+			layers & MLX5_FLOW_LAYER_OUTER_L4 ? "l4" : "-",
+			layers & MLX5_FLOW_LAYER_TUNNEL ? "T" : "-",
+			layers & MLX5_FLOW_LAYER_INNER_L2 ? "l2" : "-",
+			layers & MLX5_FLOW_LAYER_INNER_L3 ? "l3" : "-",
+			layers & MLX5_FLOW_LAYER_INNER_L4 ? "l4" : "-");
 		if (verbs->attr) {
 			struct ibv_spec_header *hdr =
 				(struct ibv_spec_header *)verbs->specs;
@@ -634,16 +665,18 @@ mlx5_flow_item_eth(const struct rte_flow_item *item, struct rte_flow *flow,
 		.src.addr_bytes = "\xff\xff\xff\xff\xff\xff",
 		.type = RTE_BE16(0xffff),
 	};
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	const unsigned int size = sizeof(struct ibv_flow_spec_eth);
 	struct ibv_flow_spec_eth eth = {
-		.type = IBV_FLOW_SPEC_ETH,
+		.type = IBV_FLOW_SPEC_ETH | (tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
-	const uint32_t layers = mlx5_flow_layers(flow);
 	int ret;
 
 	if (!flow->expand) {
-		if (layers & MLX5_FLOW_LAYER_OUTER_L2)
+		if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L2 :
+			      MLX5_FLOW_LAYER_OUTER_L2))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
@@ -658,7 +691,8 @@ mlx5_flow_item_eth(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (ret)
 			return ret;
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L2);
+	mlx5_flow_layers_update(flow, tunnel ? MLX5_FLOW_LAYER_INNER_L2 :
+				MLX5_FLOW_LAYER_OUTER_L2);
 	if (size > flow_size)
 		return size;
 	if (spec) {
@@ -694,7 +728,7 @@ mlx5_flow_item_vlan_update(struct ibv_flow_attr *attr,
 			   struct ibv_flow_spec_eth *eth)
 {
 	unsigned int i;
-	enum ibv_flow_spec_type search = IBV_FLOW_SPEC_ETH;
+	const enum ibv_flow_spec_type search = eth->type;
 	struct ibv_spec_header *hdr = (struct ibv_spec_header *)
 		((uint8_t *)attr + sizeof(struct ibv_flow_attr));
 
@@ -737,17 +771,20 @@ mlx5_flow_item_vlan(const struct rte_flow_item *item, struct rte_flow *flow,
 		.tci = RTE_BE16(0x0fff),
 	};
 	unsigned int size = sizeof(struct ibv_flow_spec_eth);
-	struct mlx5_flow_verbs *verbs = flow->cur_verbs;
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	struct ibv_flow_spec_eth eth = {
-		.type = IBV_FLOW_SPEC_ETH,
+		.type = IBV_FLOW_SPEC_ETH | (tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
 	int ret;
-	const uint32_t lm = MLX5_FLOW_LAYER_OUTER_L3 |
-			MLX5_FLOW_LAYER_OUTER_L4;
-	const uint32_t vlanm = MLX5_FLOW_LAYER_OUTER_VLAN;
-	const uint32_t l2m = MLX5_FLOW_LAYER_OUTER_L2;
-	const uint32_t layers = mlx5_flow_layers(flow);
+	const uint32_t lm = tunnel ? (MLX5_FLOW_LAYER_INNER_L3 |
+				      MLX5_FLOW_LAYER_INNER_L4) :
+		(MLX5_FLOW_LAYER_OUTER_L3 | MLX5_FLOW_LAYER_OUTER_L4);
+	const uint32_t vlanm = tunnel ? MLX5_FLOW_LAYER_INNER_VLAN :
+		MLX5_FLOW_LAYER_OUTER_VLAN;
+	const uint32_t l2m = tunnel ? MLX5_FLOW_LAYER_INNER_L2 :
+		MLX5_FLOW_LAYER_OUTER_L2;
 
 	if (!flow->expand) {
 		if (layers & vlanm)
@@ -801,12 +838,17 @@ mlx5_flow_item_vlan(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (size <= flow_size)
 			mlx5_flow_spec_verbs_add(flow, &eth, size);
 	} else {
-		if (verbs->attr)
-			mlx5_flow_item_vlan_update(verbs->attr, &eth);
+		if (flow->cur_verbs)
+			mlx5_flow_item_vlan_update(flow->cur_verbs->attr,
+						   &eth);
 		size = 0; /**< Only an update is done in eth specification. */
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L2 |
-				MLX5_FLOW_LAYER_OUTER_VLAN);
+	mlx5_flow_layers_update(flow,
+				tunnel ?
+				(MLX5_FLOW_LAYER_INNER_L2 |
+				 MLX5_FLOW_LAYER_INNER_VLAN) :
+				(MLX5_FLOW_LAYER_OUTER_L2 |
+				 MLX5_FLOW_LAYER_OUTER_VLAN));
 	return size;
 }
 
@@ -840,22 +882,26 @@ mlx5_flow_item_ipv4(const struct rte_flow_item *item, struct rte_flow *flow,
 			.next_proto_id = 0xff,
 		},
 	};
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	unsigned int size = sizeof(struct ibv_flow_spec_ipv4_ext);
 	struct ibv_flow_spec_ipv4_ext ipv4 = {
-		.type = IBV_FLOW_SPEC_IPV4_EXT,
+		.type = IBV_FLOW_SPEC_IPV4_EXT |
+			(tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
 	int ret;
-	const uint32_t layers = mlx5_flow_layers(flow);
 
 	if (!flow->expand) {
-		if (layers & MLX5_FLOW_LAYER_OUTER_L3)
+		if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L3 :
+			      MLX5_FLOW_LAYER_OUTER_L3))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
 						  "multiple L3 layers not"
 						  " supported");
-		else if (layers & MLX5_FLOW_LAYER_OUTER_L4)
+		else if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L4 :
+				   MLX5_FLOW_LAYER_OUTER_L4))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
@@ -870,7 +916,9 @@ mlx5_flow_item_ipv4(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (ret < 0)
 			return ret;
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L3_IPV4);
+	mlx5_flow_layers_update(flow,
+				tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV4 :
+				MLX5_FLOW_LAYER_OUTER_L3_IPV4);
 	if (size > flow_size)
 		return size;
 	if (spec) {
@@ -931,22 +979,25 @@ mlx5_flow_item_ipv6(const struct rte_flow_item *item, struct rte_flow *flow,
 			.hop_limits = 0xff,
 		},
 	};
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	unsigned int size = sizeof(struct ibv_flow_spec_ipv6);
 	struct ibv_flow_spec_ipv6 ipv6 = {
-		.type = IBV_FLOW_SPEC_IPV6,
+		.type = IBV_FLOW_SPEC_IPV6 | (tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
 	int ret;
-	const uint32_t layers = mlx5_flow_layers(flow);
 
 	if (!flow->expand) {
-		if (layers & MLX5_FLOW_LAYER_OUTER_L3)
+		if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L3 :
+			      MLX5_FLOW_LAYER_OUTER_L3))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
 						  "multiple L3 layers not"
 						  " supported");
-		else if (layers & MLX5_FLOW_LAYER_OUTER_L4)
+		else if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L4 :
+				   MLX5_FLOW_LAYER_OUTER_L4))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
@@ -961,7 +1012,9 @@ mlx5_flow_item_ipv6(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (ret < 0)
 			return ret;
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L3_IPV6);
+	mlx5_flow_layers_update(flow,
+				tunnel ? MLX5_FLOW_LAYER_INNER_L3_IPV6 :
+				MLX5_FLOW_LAYER_OUTER_L3_IPV6);
 	if (size > flow_size)
 		return size;
 	if (spec) {
@@ -1029,22 +1082,25 @@ mlx5_flow_item_udp(const struct rte_flow_item *item, struct rte_flow *flow,
 {
 	const struct rte_flow_item_udp *spec = item->spec;
 	const struct rte_flow_item_udp *mask = item->mask;
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	unsigned int size = sizeof(struct ibv_flow_spec_tcp_udp);
 	struct ibv_flow_spec_tcp_udp udp = {
-		.type = IBV_FLOW_SPEC_UDP,
+		.type = IBV_FLOW_SPEC_UDP | (tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
 	int ret;
-	const uint32_t layers = mlx5_flow_layers(flow);
 
 	if (!flow->expand) {
-		if (!(layers & MLX5_FLOW_LAYER_OUTER_L3))
+		if (!(layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L3 :
+				MLX5_FLOW_LAYER_OUTER_L3)))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
 						  "L3 is mandatory to filter"
 						  " on L4");
-		if (layers & MLX5_FLOW_LAYER_OUTER_L4)
+		if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L4 :
+			      MLX5_FLOW_LAYER_OUTER_L4))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
@@ -1059,7 +1115,9 @@ mlx5_flow_item_udp(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (ret < 0)
 			return ret;
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L4_UDP);
+	mlx5_flow_layers_update(flow,
+				tunnel ? MLX5_FLOW_LAYER_INNER_L4_UDP :
+				MLX5_FLOW_LAYER_OUTER_L4_UDP);
 	if (size > flow_size)
 		return size;
 	if (spec) {
@@ -1097,22 +1155,25 @@ mlx5_flow_item_tcp(const struct rte_flow_item *item, struct rte_flow *flow,
 {
 	const struct rte_flow_item_tcp *spec = item->spec;
 	const struct rte_flow_item_tcp *mask = item->mask;
+	const uint32_t layers = mlx5_flow_layers(flow);
+	const int tunnel = !!(layers & MLX5_FLOW_LAYER_TUNNEL);
 	unsigned int size = sizeof(struct ibv_flow_spec_tcp_udp);
 	struct ibv_flow_spec_tcp_udp tcp = {
-		.type = IBV_FLOW_SPEC_TCP,
+		.type = IBV_FLOW_SPEC_TCP | (tunnel ? IBV_FLOW_SPEC_INNER : 0),
 		.size = size,
 	};
 	int ret;
-	const uint32_t layers = mlx5_flow_layers(flow);
 
 	if (!flow->expand) {
-		if (!(layers & MLX5_FLOW_LAYER_OUTER_L3))
+		if (!(layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L3 :
+				MLX5_FLOW_LAYER_OUTER_L3)))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
 						  "L3 is mandatory to filter"
 						  " on L4");
-		if (layers & MLX5_FLOW_LAYER_OUTER_L4)
+		if (layers & (tunnel ? MLX5_FLOW_LAYER_INNER_L4 :
+			      MLX5_FLOW_LAYER_OUTER_L4))
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
 						  item,
@@ -1127,7 +1188,9 @@ mlx5_flow_item_tcp(const struct rte_flow_item *item, struct rte_flow *flow,
 		if (ret < 0)
 			return ret;
 	}
-	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_OUTER_L4_TCP);
+	mlx5_flow_layers_update(flow,
+				tunnel ? MLX5_FLOW_LAYER_INNER_L4_TCP :
+				MLX5_FLOW_LAYER_OUTER_L4_TCP);
 	if (size > flow_size)
 		return size;
 	if (spec) {
@@ -1143,6 +1206,102 @@ mlx5_flow_item_tcp(const struct rte_flow_item *item, struct rte_flow *flow,
 	return size;
 }
 
+/**
+ * Validate VXLAN layer and possibly create the Verbs specification.
+ *
+ * @param item[in]
+ *   Item specification.
+ * @param flow[in, out]
+ *   Pointer to flow structure.
+ * @param flow_size[in]
+ *   Size in bytes of the available space for to store the flow information.
+ * @param error
+ *   Pointer to error structure.
+ *
+ * @return
+ *   size in bytes necessary for the conversion, a negative errno value
+ *   otherwise and rte_errno is set.
+ */
+static int
+mlx5_flow_item_vxlan(const struct rte_flow_item *item, struct rte_flow *flow,
+		     const size_t flow_size, struct rte_flow_error *error)
+{
+	const struct rte_flow_item_vxlan *spec = item->spec;
+	const struct rte_flow_item_vxlan *mask = item->mask;
+	const uint32_t layers = mlx5_flow_layers(flow);
+	unsigned int size = sizeof(struct ibv_flow_spec_tunnel);
+	struct ibv_flow_spec_tunnel vxlan = {
+		.type = IBV_FLOW_SPEC_VXLAN_TUNNEL,
+		.size = size,
+	};
+	int ret;
+	union vni {
+		uint32_t vlan_id;
+		uint8_t vni[4];
+	} id = { .vlan_id = 0, };
+
+	if (!flow->expand) {
+		if (layers & MLX5_FLOW_LAYER_TUNNEL)
+			return rte_flow_error_set(error, ENOTSUP,
+						  RTE_FLOW_ERROR_TYPE_ITEM,
+						  item,
+						  "a tunnel is already"
+						  " present");
+		/*
+		 * Verify only UDPv4 is present as defined in
+		 * https://tools.ietf.org/html/rfc7348
+		 */
+		if (!(layers & MLX5_FLOW_LAYER_OUTER_L4_UDP))
+			return rte_flow_error_set(error, ENOTSUP,
+						  RTE_FLOW_ERROR_TYPE_ITEM,
+						  item,
+						  "no outer UDP layer found");
+		if (!mask)
+			mask = &rte_flow_item_vxlan_mask;
+		ret = mlx5_flow_item_validate
+			(item, (const uint8_t *)mask,
+			 (const uint8_t *)&rte_flow_item_vxlan_mask,
+			 sizeof(struct rte_flow_item_vxlan), error);
+		if (ret < 0)
+			return ret;
+	}
+	if (spec) {
+		memcpy(&id.vni[1], spec->vni, 3);
+		vxlan.val.tunnel_id = id.vlan_id;
+		memcpy(&id.vni[1], mask->vni, 3);
+		vxlan.mask.tunnel_id = id.vlan_id;
+		/* Remove unwanted bits from values. */
+		vxlan.val.tunnel_id &= vxlan.mask.tunnel_id;
+	}
+	if (!flow->expand) {
+		/*
+		 * Tunnel id 0 is equivalent as not adding a VXLAN layer, if
+		 * only this layer is defined in the Verbs specification it is
+		 * interpreted as wildcard and all packets will match this
+		 * rule, if it follows a full stack layer (ex: eth / ipv4 /
+		 * udp), all packets matching the layers before will also
+		 * match this rule.  To avoid such situation, VNI 0 is
+		 * currently refused.
+		 */
+		if (!vxlan.val.tunnel_id)
+			return rte_flow_error_set(error, EINVAL,
+						  RTE_FLOW_ERROR_TYPE_ITEM,
+						  item,
+						  "VXLAN vni cannot be 0");
+		if (!(layers & MLX5_FLOW_LAYER_OUTER))
+			return rte_flow_error_set(error, EINVAL,
+						  RTE_FLOW_ERROR_TYPE_ITEM,
+						  item,
+						  "VXLAN tunnel must be fully"
+						  " defined");
+	}
+	if (size <= flow_size)
+		mlx5_flow_spec_verbs_add(flow, &vxlan, size);
+	mlx5_flow_layers_update(flow, MLX5_FLOW_LAYER_VXLAN);
+	flow->ptype = RTE_PTYPE_TUNNEL_VXLAN | RTE_PTYPE_L4_UDP;
+	return size;
+}
+
 /**
  * Validate items provided by the user.
  *
@@ -1191,6 +1350,9 @@ mlx5_flow_items(const struct rte_flow_item items[],
 		case RTE_FLOW_ITEM_TYPE_TCP:
 			ret = mlx5_flow_item_tcp(items, flow, remain, error);
 			break;
+		case RTE_FLOW_ITEM_TYPE_VXLAN:
+			ret = mlx5_flow_item_vxlan(items, flow, remain, error);
+			break;
 		default:
 			return rte_flow_error_set(error, ENOTSUP,
 						  RTE_FLOW_ERROR_TYPE_ITEM,
-- 
2.18.0