net/mlx5: add flow UDP item
authorNelio Laranjeiro <nelio.laranjeiro@6wind.com>
Thu, 12 Jul 2018 09:30:56 +0000 (11:30 +0200)
committerShahaf Shuler <shahafs@mellanox.com>
Thu, 12 Jul 2018 10:10:03 +0000 (12:10 +0200)
Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
drivers/net/mlx5/mlx5_flow.c

index 513f70d..0096ed8 100644 (file)
@@ -52,6 +52,9 @@ extern const struct eth_dev_ops mlx5_dev_ops_isolate;
 #define MLX5_FLOW_FATE_DROP (1u << 0)
 #define MLX5_FLOW_FATE_QUEUE (1u << 1)
 
+/* possible L3 layers protocols filtering. */
+#define MLX5_IP_PROTOCOL_UDP 17
+
 /** Handles information leading to a drop fate. */
 struct mlx5_flow_verbs {
        unsigned int size; /**< Size of the attribute. */
@@ -68,10 +71,12 @@ struct mlx5_flow_verbs {
 struct rte_flow {
        TAILQ_ENTRY(rte_flow) next; /**< Pointer to the next flow structure. */
        struct rte_flow_attr attributes; /**< User flow attribute. */
+       uint32_t l3_protocol_en:1; /**< Protocol filtering requested. */
        uint32_t layers;
        /**< Bit-fields of present layers see MLX5_FLOW_LAYER_*. */
        uint32_t fate;
        /**< Bit-fields of present fate see MLX5_FLOW_FATE_*. */
+       uint8_t l3_protocol; /**< valid when l3_protocol_en is set. */
        struct mlx5_flow_verbs verbs; /* Verbs flow. */
        uint16_t queue; /**< Destination queue to redirect traffic to. */
 };
@@ -568,8 +573,6 @@ mlx5_flow_item_ipv4(const struct rte_flow_item *item, struct rte_flow *flow,
        if (ret < 0)
                return ret;
        flow->layers |= MLX5_FLOW_LAYER_OUTER_L3_IPV4;
-       if (size > flow_size)
-               return size;
        if (spec) {
                ipv4.val = (struct ibv_flow_ipv4_ext_filter){
                        .src_ip = spec->hdr.src_addr,
@@ -589,7 +592,10 @@ mlx5_flow_item_ipv4(const struct rte_flow_item *item, struct rte_flow *flow,
                ipv4.val.proto &= ipv4.mask.proto;
                ipv4.val.tos &= ipv4.mask.tos;
        }
-       mlx5_flow_spec_verbs_add(flow, &ipv4, size);
+       flow->l3_protocol_en = !!ipv4.mask.proto;
+       flow->l3_protocol = ipv4.val.proto;
+       if (size <= flow_size)
+               mlx5_flow_spec_verbs_add(flow, &ipv4, size);
        return size;
 }
 
@@ -660,8 +666,6 @@ mlx5_flow_item_ipv6(const struct rte_flow_item *item, struct rte_flow *flow,
        if (ret < 0)
                return ret;
        flow->layers |= MLX5_FLOW_LAYER_OUTER_L3_IPV6;
-       if (size > flow_size)
-               return size;
        if (spec) {
                unsigned int i;
                uint32_t vtc_flow_val;
@@ -701,7 +705,85 @@ mlx5_flow_item_ipv6(const struct rte_flow_item *item, struct rte_flow *flow,
                ipv6.val.next_hdr &= ipv6.mask.next_hdr;
                ipv6.val.hop_limit &= ipv6.mask.hop_limit;
        }
-       mlx5_flow_spec_verbs_add(flow, &ipv6, size);
+       flow->l3_protocol_en = !!ipv6.mask.next_hdr;
+       flow->l3_protocol = ipv6.val.next_hdr;
+       if (size <= flow_size)
+               mlx5_flow_spec_verbs_add(flow, &ipv6, size);
+       return size;
+}
+
+/**
+ * Convert the @p item into a Verbs specification after ensuring the NIC
+ * will understand and process it correctly.
+ * If the necessary size for the conversion is greater than the @p flow_size,
+ * nothing is written in @p flow, the validation is still performed.
+ *
+ * @param[in] item
+ *   Item specification.
+ * @param[in, out] flow
+ *   Pointer to flow structure.
+ * @param[in] flow_size
+ *   Size in bytes of the available space in @p flow, if too small, nothing is
+ *   written.
+ * @param[out] error
+ *   Pointer to error structure.
+ *
+ * @return
+ *   On success the number of bytes consumed/necessary, if the returned value
+ *   is lesser or equal to @p flow_size, the @p item has fully been converted,
+ *   otherwise another call with this returned memory size should be done.
+ *   On error, a negative errno value is returned and rte_errno is set.
+ */
+static int
+mlx5_flow_item_udp(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_udp *spec = item->spec;
+       const struct rte_flow_item_udp *mask = item->mask;
+       unsigned int size = sizeof(struct ibv_flow_spec_tcp_udp);
+       struct ibv_flow_spec_tcp_udp udp = {
+               .type = IBV_FLOW_SPEC_UDP,
+               .size = size,
+       };
+       int ret;
+
+       if (!(flow->layers & 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 (flow->layers & MLX5_FLOW_LAYER_OUTER_L4)
+               return rte_flow_error_set(error, ENOTSUP,
+                                         RTE_FLOW_ERROR_TYPE_ITEM,
+                                         item,
+                                         "L4 layer is already present");
+       if (flow->l3_protocol_en && flow->l3_protocol != MLX5_IP_PROTOCOL_UDP)
+               return rte_flow_error_set(error, ENOTSUP,
+                                         RTE_FLOW_ERROR_TYPE_ITEM,
+                                         item,
+                                         "protocol filtering not compatible"
+                                         " with UDP layer");
+       if (!mask)
+               mask = &rte_flow_item_udp_mask;
+       ret = mlx5_flow_item_acceptable
+               (item, (const uint8_t *)mask,
+                (const uint8_t *)&rte_flow_item_udp_mask,
+                sizeof(struct rte_flow_item_udp), error);
+       if (ret < 0)
+               return ret;
+       flow->layers |= MLX5_FLOW_LAYER_OUTER_L4_UDP;
+       if (size > flow_size)
+               return size;
+       if (spec) {
+               udp.val.dst_port = spec->hdr.dst_port;
+               udp.val.src_port = spec->hdr.src_port;
+               udp.mask.dst_port = mask->hdr.dst_port;
+               udp.mask.src_port = mask->hdr.src_port;
+               /* Remove unwanted bits from values. */
+               udp.val.src_port &= udp.mask.src_port;
+               udp.val.dst_port &= udp.mask.dst_port;
+       }
+       mlx5_flow_spec_verbs_add(flow, &udp, size);
        return size;
 }
 
@@ -756,6 +838,9 @@ mlx5_flow_items(const struct rte_flow_item pattern[],
                case RTE_FLOW_ITEM_TYPE_IPV6:
                        ret = mlx5_flow_item_ipv6(pattern, flow, remain, error);
                        break;
+               case RTE_FLOW_ITEM_TYPE_UDP:
+                       ret = mlx5_flow_item_udp(pattern, flow, remain, error);
+                       break;
                default:
                        return rte_flow_error_set(error, ENOTSUP,
                                                  RTE_FLOW_ERROR_TYPE_ITEM,