[dpdk-dev] [PATCH 2/2] net/mlx5: add flow match on GENEVE item

[Moti Haimovsky <motih@mellanox.com>]

This commit adds support for matching flows on Geneve headers.

Signed-off-by: Moti Haimovsky <motih@mellanox.com>
---
 doc/guides/nics/mlx5.rst               | 10 +++-
 doc/guides/rel_notes/release_19_11.rst |  1 +
 drivers/net/mlx5/mlx5.h                |  2 +
 drivers/net/mlx5/mlx5_devx_cmds.c      |  6 +++
 drivers/net/mlx5/mlx5_flow.c           | 89 ++++++++++++++++++++++++++++++++++
 drivers/net/mlx5/mlx5_flow.h           | 38 ++++++++++++++-
 drivers/net/mlx5/mlx5_flow_dv.c        | 84 ++++++++++++++++++++++++++++++++
 drivers/net/mlx5/mlx5_prm.h            |  9 +++-
 8 files changed, 235 insertions(+), 4 deletions(-)

diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst
index 414c9c1..08039bc 100644
--- a/doc/guides/nics/mlx5.rst
+++ b/doc/guides/nics/mlx5.rst
@@ -87,7 +87,7 @@ Features
 - RX interrupts.
 - Statistics query including Basic, Extended and per queue.
 - Rx HW timestamp.
-- Tunnel types: VXLAN, L3 VXLAN, VXLAN-GPE, GRE, MPLSoGRE, MPLSoUDP, IP-in-IP.
+- Tunnel types: VXLAN, L3 VXLAN, VXLAN-GPE, GRE, MPLSoGRE, MPLSoUDP, IP-in-IP, Geneve.
 - Tunnel HW offloads: packet type, inner/outer RSS, IP and UDP checksum verification.
 - NIC HW offloads: encapsulation (vxlan, gre, mplsoudp, mplsogre), NAT, routing, TTL
   increment/decrement, count, drop, mark. For details please see :ref:`mlx5_offloads_support`.
@@ -146,6 +146,14 @@ Limitations
 
 - L3 VXLAN and VXLAN-GPE tunnels cannot be supported together with MPLSoGRE and MPLSoUDP.
 
+- Match on Geneve header supports the following fields only:
+
+     - VNI
+     - OAM
+     - protocol type
+     - options length
+       Currently, the only supported options length value is 0.
+
 - VF: flow rules created on VF devices can only match traffic targeted at the
   configured MAC addresses (see ``rte_eth_dev_mac_addr_add()``).
 
diff --git a/doc/guides/rel_notes/release_19_11.rst b/doc/guides/rel_notes/release_19_11.rst
index cd4e350..a903533 100644
--- a/doc/guides/rel_notes/release_19_11.rst
+++ b/doc/guides/rel_notes/release_19_11.rst
@@ -286,4 +286,5 @@ Tested Platforms
   * Added support for VLAN push flow offload command.
   * Added support for VLAN set PCP offload command.
   * Added support for VLAN set VID offload command.
+  * Added support for matching on packets withe Geneve tunnel header.
 
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 024a403..bac1c0a 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -179,6 +179,8 @@ struct mlx5_hca_attr {
 	uint32_t wqe_vlan_insert:1;
 	uint32_t wqe_inline_mode:2;
 	uint32_t vport_inline_mode:3;
+	uint32_t tunnel_stateless_geneve_rx:1;
+	uint32_t geneve_max_opt_len:1; /* 0x0: 14DW, 0x1: 63DW */
 	uint32_t lro_cap:1;
 	uint32_t tunnel_lro_gre:1;
 	uint32_t tunnel_lro_vxlan:1;
diff --git a/drivers/net/mlx5/mlx5_devx_cmds.c b/drivers/net/mlx5/mlx5_devx_cmds.c
index 01e4094..51947d3 100644
--- a/drivers/net/mlx5/mlx5_devx_cmds.c
+++ b/drivers/net/mlx5/mlx5_devx_cmds.c
@@ -376,6 +376,12 @@ struct mlx5_devx_obj *
 			MLX5_GET(per_protocol_networking_offload_caps, hcattr,
 				 lro_timer_supported_periods[i]);
 	}
+	attr->tunnel_stateless_geneve_rx =
+			    MLX5_GET(per_protocol_networking_offload_caps,
+				     hcattr, tunnel_stateless_geneve_rx);
+	attr->geneve_max_opt_len =
+		    MLX5_GET(per_protocol_networking_offload_caps,
+			     hcattr, max_geneve_opt_len);
 	attr->wqe_inline_mode = MLX5_GET(per_protocol_networking_offload_caps,
 					 hcattr, wqe_inline_mode);
 	if (attr->wqe_inline_mode != MLX5_CAP_INLINE_MODE_VPORT_CONTEXT)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 578d003..d4d956f 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -1913,6 +1913,95 @@ uint32_t mlx5_flow_adjust_priority(struct rte_eth_dev *dev, int32_t priority,
 }
 
 /**
+ * Validate Geneve item.
+ *
+ * @param[in] item
+ *   Item specification.
+ * @param[in] itemFlags
+ *   Bit-fields that holds the items detected until now.
+ * @param[in] enPriv
+ *   Pointer to the private data structure.
+ * @param[out] error
+ *   Pointer to error structure.
+ *
+ * @return
+ *   0 on success, a negative errno value otherwise and rte_errno is set.
+ */
+
+int
+mlx5_flow_validate_item_geneve(const struct rte_flow_item *item,
+			       uint64_t item_flags,
+			       struct rte_eth_dev *dev,
+			       struct rte_flow_error *error)
+{
+	struct mlx5_priv *priv = dev->data->dev_private;
+	const struct rte_flow_item_geneve *spec = item->spec;
+	const struct rte_flow_item_geneve *mask = item->mask;
+	int ret;
+	uint16_t gbhdr;
+	uint8_t opt_len = priv->config.hca_attr.geneve_max_opt_len ?
+			  MLX5_GENEVE_OPT_LEN_1 : MLX5_GENEVE_OPT_LEN_0;
+	const struct rte_flow_item_geneve nic_mask = {
+		.ver_opt_len_o_c_rsvd0 = RTE_BE16(0x3f80),
+		.vni = "\xff\xff\xff",
+		.protocol = RTE_BE16(UINT16_MAX),
+	};
+
+	if (!(priv->config.hca_attr.flex_parser_protocols &
+	      MLX5_HCA_FLEX_GENEVE_ENABLED) ||
+	    !priv->config.hca_attr.tunnel_stateless_geneve_rx)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ITEM, item,
+					  "L3 Geneve is not enabled by device"
+					  " parameter and/or not configured in"
+					  " firmware");
+	if (item_flags & MLX5_FLOW_LAYER_TUNNEL)
+		return rte_flow_error_set(error, ENOTSUP,
+					  RTE_FLOW_ERROR_TYPE_ITEM, item,
+					  "multiple tunnel layers not"
+					  " supported");
+	/*
+	 * Verify only UDPv4 is present as defined in
+	 * https://tools.ietf.org/html/rfc7348
+	 */
+	if (!(item_flags & MLX5_FLOW_LAYER_OUTER_L4_UDP))
+		return rte_flow_error_set(error, EINVAL,
+					  RTE_FLOW_ERROR_TYPE_ITEM, item,
+					  "no outer UDP layer found");
+	if (!mask)
+		mask = &rte_flow_item_geneve_mask;
+	ret = mlx5_flow_item_acceptable
+				  (item, (const uint8_t *)mask,
+				   (const uint8_t *)&nic_mask,
+				   sizeof(struct rte_flow_item_geneve), error);
+	if (ret)
+		return ret;
+	if (spec) {
+		gbhdr = rte_be_to_cpu_16(spec->ver_opt_len_o_c_rsvd0);
+		if (MLX5_GENEVE_VER_VAL(gbhdr) ||
+		     MLX5_GENEVE_CRITO_VAL(gbhdr) ||
+		     MLX5_GENEVE_RSVD_VAL(gbhdr) || spec->rsvd1)
+			return rte_flow_error_set(error, ENOTSUP,
+						  RTE_FLOW_ERROR_TYPE_ITEM,
+						  item,
+						  "Geneve protocol unsupported"
+						  " fields are being used");
+		if (MLX5_GENEVE_OPTLEN_VAL(gbhdr) > opt_len)
+			return rte_flow_error_set
+					(error, ENOTSUP,
+					 RTE_FLOW_ERROR_TYPE_ITEM,
+					 item,
+					 "Unsupported Geneve options length");
+	}
+	if (!(item_flags & MLX5_FLOW_LAYER_OUTER))
+		return rte_flow_error_set
+				    (error, ENOTSUP,
+				     RTE_FLOW_ERROR_TYPE_ITEM, item,
+				     "Geneve tunnel must be fully defined");
+	return 0;
+}
+
+/**
  * Validate MPLS item.
  *
  * @param[in] dev
diff --git a/drivers/net/mlx5/mlx5_flow.h b/drivers/net/mlx5/mlx5_flow.h
index 235bccd..9658db1 100644
--- a/drivers/net/mlx5/mlx5_flow.h
+++ b/drivers/net/mlx5/mlx5_flow.h
@@ -63,6 +63,7 @@
 #define MLX5_FLOW_LAYER_IPIP (1u << 21)
 #define MLX5_FLOW_LAYER_IPV6_ENCAP (1u << 22)
 #define MLX5_FLOW_LAYER_NVGRE (1u << 23)
+#define MLX5_FLOW_LAYER_GENEVE (1u << 24)
 
 /* Outer Masks. */
 #define MLX5_FLOW_LAYER_OUTER_L3 \
@@ -83,7 +84,8 @@
 #define MLX5_FLOW_LAYER_TUNNEL \
 	(MLX5_FLOW_LAYER_VXLAN | MLX5_FLOW_LAYER_VXLAN_GPE | \
 	 MLX5_FLOW_LAYER_GRE | MLX5_FLOW_LAYER_NVGRE | MLX5_FLOW_LAYER_MPLS | \
-	 MLX5_FLOW_LAYER_IPIP | MLX5_FLOW_LAYER_IPV6_ENCAP)
+	 MLX5_FLOW_LAYER_IPIP | MLX5_FLOW_LAYER_IPV6_ENCAP | \
+	 MLX5_FLOW_LAYER_GENEVE)
 
 /* Inner Masks. */
 #define MLX5_FLOW_LAYER_INNER_L3 \
@@ -188,6 +190,9 @@
 #define MLX5_UDP_PORT_VXLAN 4789
 #define MLX5_UDP_PORT_VXLAN_GPE 4790
 
+/* UDP port numbers for GENEVE. */
+#define MLX5_UDP_PORT_GENEVE 6081
+
 /* Priority reserved for default flows. */
 #define MLX5_FLOW_PRIO_RSVD ((uint32_t)-1)
 
@@ -220,6 +225,33 @@
 /* IBV hash source bits  for IPV6. */
 #define MLX5_IPV6_IBV_RX_HASH (IBV_RX_HASH_SRC_IPV6 | IBV_RX_HASH_DST_IPV6)
 
+
+/* Geneve header first 16Bit */
+#define MLX5_GENEVE_VER_MASK 0x3
+#define MLX5_GENEVE_VER_SHIFT 14
+#define MLX5_GENEVE_VER_VAL(a) \
+		(((a) >> (MLX5_GENEVE_VER_SHIFT)) & (MLX5_GENEVE_VER_MASK))
+#define MLX5_GENEVE_OPTLEN_MASK 0x3F
+#define MLX5_GENEVE_OPTLEN_SHIFT 7
+#define MLX5_GENEVE_OPTLEN_VAL(a) \
+	    (((a) >> (MLX5_GENEVE_OPTLEN_SHIFT)) & (MLX5_GENEVE_OPTLEN_MASK))
+#define MLX5_GENEVE_OAMF_MASK 0x1
+#define MLX5_GENEVE_OAMF_SHIFT 7
+#define MLX5_GENEVE_OAMF_VAL(a) \
+		(((a) >> (MLX5_GENEVE_OAMF_SHIFT)) & (MLX5_GENEVE_OAMF_MASK))
+#define MLX5_GENEVE_CRITO_MASK 0x1
+#define MLX5_GENEVE_CRITO_SHIFT 6
+#define MLX5_GENEVE_CRITO_VAL(a) \
+		(((a) >> (MLX5_GENEVE_CRITO_SHIFT)) & (MLX5_GENEVE_CRITO_MASK))
+#define MLX5_GENEVE_RSVD_MASK 0x3F
+#define MLX5_GENEVE_RSVD_VAL(a) ((a) & (MLX5_GENEVE_RSVD_MASK))
+/*
+ * The length of the Geneve options fields, expressed in four byte multiples,
+ * not including the eight byte fixed tunnel.
+ */
+#define MLX5_GENEVE_OPT_LEN_0 14
+#define MLX5_GENEVE_OPT_LEN_1 63
+
 enum mlx5_flow_drv_type {
 	MLX5_FLOW_TYPE_MIN,
 	MLX5_FLOW_TYPE_DV,
@@ -556,4 +588,8 @@ int mlx5_flow_validate_item_nvgre(const struct rte_flow_item *item,
 				  uint64_t item_flags,
 				  uint8_t target_protocol,
 				  struct rte_flow_error *error);
+int mlx5_flow_validate_item_geneve(const struct rte_flow_item *item,
+				   uint64_t item_flags,
+				   struct rte_eth_dev *dev,
+				   struct rte_flow_error *error);
 #endif /* RTE_PMD_MLX5_FLOW_H_ */
diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c
index f0422dc..b1aa427 100644
--- a/drivers/net/mlx5/mlx5_flow_dv.c
+++ b/drivers/net/mlx5/mlx5_flow_dv.c
@@ -3434,6 +3434,14 @@ struct field_modify_info modify_tcp[] = {
 				return ret;
 			last_item = MLX5_FLOW_LAYER_VXLAN_GPE;
 			break;
+		case RTE_FLOW_ITEM_TYPE_GENEVE:
+			ret = mlx5_flow_validate_item_geneve(items,
+							     item_flags, dev,
+							     error);
+			if (ret < 0)
+				return ret;
+			last_item = MLX5_FLOW_LAYER_VXLAN_GPE;
+			break;
 		case RTE_FLOW_ITEM_TYPE_MPLS:
 			ret = mlx5_flow_validate_item_mpls(dev, items,
 							   item_flags,
@@ -4492,6 +4500,77 @@ struct field_modify_info modify_tcp[] = {
 }
 
 /**
+ * Add Geneve item to matcher and to the value.
+ *
+ * @param[in, out] matcher
+ *   Flow matcher.
+ * @param[in, out] key
+ *   Flow matcher value.
+ * @param[in] item
+ *   Flow pattern to translate.
+ * @param[in] inner
+ *   Item is inner pattern.
+ */
+
+static void
+flow_dv_translate_item_geneve(void *matcher, void *key,
+			      const struct rte_flow_item *item, int inner)
+{
+	const struct rte_flow_item_geneve *geneve_m = item->mask;
+	const struct rte_flow_item_geneve *geneve_v = item->spec;
+	void *headers_m;
+	void *headers_v;
+	void *misc_m = MLX5_ADDR_OF(fte_match_param, matcher, misc_parameters);
+	void *misc_v = MLX5_ADDR_OF(fte_match_param, key, misc_parameters);
+	uint16_t dport;
+	uint16_t gbhdr_m;
+	uint16_t gbhdr_v;
+	char *vni_m;
+	char *vni_v;
+	size_t size, i;
+
+	if (inner) {
+		headers_m = MLX5_ADDR_OF(fte_match_param, matcher,
+					 inner_headers);
+		headers_v = MLX5_ADDR_OF(fte_match_param, key, inner_headers);
+	} else {
+		headers_m = MLX5_ADDR_OF(fte_match_param, matcher,
+					 outer_headers);
+		headers_v = MLX5_ADDR_OF(fte_match_param, key, outer_headers);
+	}
+	dport = MLX5_UDP_PORT_GENEVE;
+	if (!MLX5_GET16(fte_match_set_lyr_2_4, headers_v, udp_dport)) {
+		MLX5_SET(fte_match_set_lyr_2_4, headers_m, udp_dport, 0xFFFF);
+		MLX5_SET(fte_match_set_lyr_2_4, headers_v, udp_dport, dport);
+	}
+	if (!geneve_v)
+		return;
+	if (!geneve_m)
+		geneve_m = &rte_flow_item_geneve_mask;
+	size = sizeof(geneve_m->vni);
+	vni_m = MLX5_ADDR_OF(fte_match_set_misc, misc_m, geneve_vni);
+	vni_v = MLX5_ADDR_OF(fte_match_set_misc, misc_v, geneve_vni);
+	memcpy(vni_m, geneve_m->vni, size);
+	for (i = 0; i < size; ++i)
+		vni_v[i] = vni_m[i] & geneve_v->vni[i];
+	MLX5_SET(fte_match_set_misc, misc_m, geneve_protocol_type,
+		 rte_be_to_cpu_16(geneve_m->protocol));
+	MLX5_SET(fte_match_set_misc, misc_v, geneve_protocol_type,
+		 rte_be_to_cpu_16(geneve_v->protocol & geneve_m->protocol));
+	gbhdr_m = rte_be_to_cpu_16(geneve_m->ver_opt_len_o_c_rsvd0);
+	gbhdr_v = rte_be_to_cpu_16(geneve_v->ver_opt_len_o_c_rsvd0);
+	MLX5_SET(fte_match_set_misc, misc_m, geneve_oam,
+		 MLX5_GENEVE_OAMF_VAL(gbhdr_m));
+	MLX5_SET(fte_match_set_misc, misc_v, geneve_oam,
+		 MLX5_GENEVE_OAMF_VAL(gbhdr_v) & MLX5_GENEVE_OAMF_VAL(gbhdr_m));
+	MLX5_SET(fte_match_set_misc, misc_m, geneve_opt_len,
+		 MLX5_GENEVE_OPTLEN_VAL(gbhdr_m));
+	MLX5_SET(fte_match_set_misc, misc_v, geneve_opt_len,
+		 MLX5_GENEVE_OPTLEN_VAL(gbhdr_v) &
+		 MLX5_GENEVE_OPTLEN_VAL(gbhdr_m));
+}
+
+/**
  * Add MPLS item to matcher and to the value.
  *
  * @param[in, out] matcher
@@ -5692,6 +5771,11 @@ struct field_modify_info modify_tcp[] = {
 						     items, tunnel);
 			last_item = MLX5_FLOW_LAYER_VXLAN_GPE;
 			break;
+		case RTE_FLOW_ITEM_TYPE_GENEVE:
+			flow_dv_translate_item_geneve(match_mask, match_value,
+						      items, tunnel);
+			last_item = MLX5_FLOW_LAYER_GENEVE;
+			break;
 		case RTE_FLOW_ITEM_TYPE_MPLS:
 			flow_dv_translate_item_mpls(match_mask, match_value,
 						    items, last_item, tunnel);
diff --git a/drivers/net/mlx5/mlx5_prm.h b/drivers/net/mlx5/mlx5_prm.h
index e429792..c86f8b8 100644
--- a/drivers/net/mlx5/mlx5_prm.h
+++ b/drivers/net/mlx5/mlx5_prm.h
@@ -553,12 +553,17 @@ struct mlx5_ifc_fte_match_set_misc_bits {
 	u8 gre_key_l[0x8];
 	u8 vxlan_vni[0x18];
 	u8 reserved_at_b8[0x8];
-	u8 reserved_at_c0[0x20];
+	u8 geneve_vni[0x18];
+	u8 reserved_at_e4[0x7];
+	u8 geneve_oam[0x1];
 	u8 reserved_at_e0[0xc];
 	u8 outer_ipv6_flow_label[0x14];
 	u8 reserved_at_100[0xc];
 	u8 inner_ipv6_flow_label[0x14];
-	u8 reserved_at_120[0xe0];
+	u8 reserved_at_120[0xa];
+	u8 geneve_opt_len[0x6];
+	u8 geneve_protocol_type[0x10];
+	u8 reserved_at_140[0xc0];
 };
 
 struct mlx5_ifc_ipv4_layout_bits {
-- 
1.8.3.1