net/mlx5: remove Netlink flow driver
authorYongseok Koh <yskoh@mellanox.com>
Mon, 24 Sep 2018 19:55:16 +0000 (19:55 +0000)
committerFerruh Yigit <ferruh.yigit@intel.com>
Thu, 11 Oct 2018 16:53:49 +0000 (18:53 +0200)
Netlink based E-Switch flow engine will be migrated to the new flow
engine.
nl_flow will be renamed to flow_tcf as it goes through Linux TC flower
interface.

Signed-off-by: Yongseok Koh <yskoh@mellanox.com>
drivers/net/mlx5/Makefile
drivers/net/mlx5/meson.build
drivers/net/mlx5/mlx5.c
drivers/net/mlx5/mlx5.h
drivers/net/mlx5/mlx5_nl_flow.c [deleted file]

index 4243b37..9c10448 100644 (file)
@@ -35,7 +35,6 @@ SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5_flow_dv.c
 SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5_flow_verbs.c
 SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5_socket.c
 SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5_nl.c
-SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += mlx5_nl_flow.c
 
 ifeq ($(CONFIG_RTE_LIBRTE_MLX5_DLOPEN_DEPS),y)
 INSTALL-$(CONFIG_RTE_LIBRTE_MLX5_PMD)-lib += $(LIB_GLUE)
index 3d09ece..e537629 100644 (file)
@@ -36,7 +36,6 @@ if build
                'mlx5_mac.c',
                'mlx5_mr.c',
                'mlx5_nl.c',
-               'mlx5_nl_flow.c',
                'mlx5_rss.c',
                'mlx5_rxmode.c',
                'mlx5_rxq.c',
index c902653..78ad40f 100644 (file)
@@ -286,8 +286,6 @@ mlx5_dev_close(struct rte_eth_dev *dev)
                close(priv->nl_socket_route);
        if (priv->nl_socket_rdma >= 0)
                close(priv->nl_socket_rdma);
-       if (priv->mnl_socket)
-               mlx5_nl_flow_socket_destroy(priv->mnl_socket);
        ret = mlx5_hrxq_ibv_verify(dev);
        if (ret)
                DRV_LOG(WARNING, "port %u some hash Rx queue still remain",
@@ -1139,34 +1137,6 @@ mlx5_dev_spawn(struct rte_device *dpdk_dev,
        claim_zero(mlx5_mac_addr_add(eth_dev, &mac, 0, 0));
        if (vf && config.vf_nl_en)
                mlx5_nl_mac_addr_sync(eth_dev);
-       priv->mnl_socket = mlx5_nl_flow_socket_create();
-       if (!priv->mnl_socket) {
-               err = -rte_errno;
-               DRV_LOG(WARNING,
-                       "flow rules relying on switch offloads will not be"
-                       " supported: cannot open libmnl socket: %s",
-                       strerror(rte_errno));
-       } else {
-               struct rte_flow_error error;
-               unsigned int ifindex = mlx5_ifindex(eth_dev);
-
-               if (!ifindex) {
-                       err = -rte_errno;
-                       error.message =
-                               "cannot retrieve network interface index";
-               } else {
-                       err = mlx5_nl_flow_init(priv->mnl_socket, ifindex,
-                                               &error);
-               }
-               if (err) {
-                       DRV_LOG(WARNING,
-                               "flow rules relying on switch offloads will"
-                               " not be supported: %s: %s",
-                               error.message, strerror(rte_errno));
-                       mlx5_nl_flow_socket_destroy(priv->mnl_socket);
-                       priv->mnl_socket = NULL;
-               }
-       }
        TAILQ_INIT(&priv->flows);
        TAILQ_INIT(&priv->ctrl_flows);
        /* Hint libmlx5 to use PMD allocator for data plane resources */
@@ -1219,8 +1189,6 @@ error:
                        close(priv->nl_socket_route);
                if (priv->nl_socket_rdma >= 0)
                        close(priv->nl_socket_rdma);
-               if (priv->mnl_socket)
-                       mlx5_nl_flow_socket_destroy(priv->mnl_socket);
                if (own_domain_id)
                        claim_zero(rte_eth_switch_domain_free(priv->domain_id));
                rte_free(priv);
index 0358ee4..4122e54 100644 (file)
@@ -158,12 +158,6 @@ struct mlx5_drop {
        struct mlx5_rxq_ibv *rxq; /* Verbs Rx queue. */
 };
 
-/** DPDK port to network interface index (ifindex) conversion. */
-struct mlx5_nl_flow_ptoi {
-       uint16_t port_id; /**< DPDK port ID. */
-       unsigned int ifindex; /**< Network interface index. */
-};
-
 struct mnl_socket;
 
 struct priv {
@@ -399,23 +393,4 @@ unsigned int mlx5_nl_ifindex(int nl, const char *name);
 int mlx5_nl_switch_info(int nl, unsigned int ifindex,
                        struct mlx5_switch_info *info);
 
-/* mlx5_nl_flow.c */
-
-int mlx5_nl_flow_transpose(void *buf,
-                          size_t size,
-                          const struct mlx5_nl_flow_ptoi *ptoi,
-                          const struct rte_flow_attr *attr,
-                          const struct rte_flow_item *pattern,
-                          const struct rte_flow_action *actions,
-                          struct rte_flow_error *error);
-void mlx5_nl_flow_brand(void *buf, uint32_t handle);
-int mlx5_nl_flow_create(struct mnl_socket *nl, void *buf,
-                       struct rte_flow_error *error);
-int mlx5_nl_flow_destroy(struct mnl_socket *nl, void *buf,
-                        struct rte_flow_error *error);
-int mlx5_nl_flow_init(struct mnl_socket *nl, unsigned int ifindex,
-                     struct rte_flow_error *error);
-struct mnl_socket *mlx5_nl_flow_socket_create(void);
-void mlx5_nl_flow_socket_destroy(struct mnl_socket *nl);
-
 #endif /* RTE_PMD_MLX5_H_ */
diff --git a/drivers/net/mlx5/mlx5_nl_flow.c b/drivers/net/mlx5/mlx5_nl_flow.c
deleted file mode 100644 (file)
index beb03c9..0000000
+++ /dev/null
@@ -1,1228 +0,0 @@
-/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright 2018 6WIND S.A.
- * Copyright 2018 Mellanox Technologies, Ltd
- */
-
-#include <assert.h>
-#include <errno.h>
-#include <libmnl/libmnl.h>
-#include <linux/if_ether.h>
-#include <linux/netlink.h>
-#include <linux/pkt_cls.h>
-#include <linux/pkt_sched.h>
-#include <linux/rtnetlink.h>
-#include <linux/tc_act/tc_gact.h>
-#include <linux/tc_act/tc_mirred.h>
-#include <netinet/in.h>
-#include <stdalign.h>
-#include <stdbool.h>
-#include <stddef.h>
-#include <stdint.h>
-#include <stdlib.h>
-#include <sys/socket.h>
-
-#include <rte_byteorder.h>
-#include <rte_errno.h>
-#include <rte_ether.h>
-#include <rte_flow.h>
-
-#include "mlx5.h"
-#include "mlx5_autoconf.h"
-
-#ifdef HAVE_TC_ACT_VLAN
-
-#include <linux/tc_act/tc_vlan.h>
-
-#else /* HAVE_TC_ACT_VLAN */
-
-#define TCA_VLAN_ACT_POP 1
-#define TCA_VLAN_ACT_PUSH 2
-#define TCA_VLAN_ACT_MODIFY 3
-#define TCA_VLAN_PARMS 2
-#define TCA_VLAN_PUSH_VLAN_ID 3
-#define TCA_VLAN_PUSH_VLAN_PROTOCOL 4
-#define TCA_VLAN_PAD 5
-#define TCA_VLAN_PUSH_VLAN_PRIORITY 6
-
-struct tc_vlan {
-       tc_gen;
-       int v_action;
-};
-
-#endif /* HAVE_TC_ACT_VLAN */
-
-/* Normally found in linux/netlink.h. */
-#ifndef NETLINK_CAP_ACK
-#define NETLINK_CAP_ACK 10
-#endif
-
-/* Normally found in linux/pkt_sched.h. */
-#ifndef TC_H_MIN_INGRESS
-#define TC_H_MIN_INGRESS 0xfff2u
-#endif
-
-/* Normally found in linux/pkt_cls.h. */
-#ifndef TCA_CLS_FLAGS_SKIP_SW
-#define TCA_CLS_FLAGS_SKIP_SW (1 << 1)
-#endif
-#ifndef HAVE_TCA_FLOWER_ACT
-#define TCA_FLOWER_ACT 3
-#endif
-#ifndef HAVE_TCA_FLOWER_FLAGS
-#define TCA_FLOWER_FLAGS 22
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_ETH_TYPE
-#define TCA_FLOWER_KEY_ETH_TYPE 8
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_ETH_DST
-#define TCA_FLOWER_KEY_ETH_DST 4
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_ETH_DST_MASK
-#define TCA_FLOWER_KEY_ETH_DST_MASK 5
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_ETH_SRC
-#define TCA_FLOWER_KEY_ETH_SRC 6
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_ETH_SRC_MASK
-#define TCA_FLOWER_KEY_ETH_SRC_MASK 7
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IP_PROTO
-#define TCA_FLOWER_KEY_IP_PROTO 9
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV4_SRC
-#define TCA_FLOWER_KEY_IPV4_SRC 10
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV4_SRC_MASK
-#define TCA_FLOWER_KEY_IPV4_SRC_MASK 11
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV4_DST
-#define TCA_FLOWER_KEY_IPV4_DST 12
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV4_DST_MASK
-#define TCA_FLOWER_KEY_IPV4_DST_MASK 13
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV6_SRC
-#define TCA_FLOWER_KEY_IPV6_SRC 14
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV6_SRC_MASK
-#define TCA_FLOWER_KEY_IPV6_SRC_MASK 15
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV6_DST
-#define TCA_FLOWER_KEY_IPV6_DST 16
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_IPV6_DST_MASK
-#define TCA_FLOWER_KEY_IPV6_DST_MASK 17
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_TCP_SRC
-#define TCA_FLOWER_KEY_TCP_SRC 18
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_TCP_SRC_MASK
-#define TCA_FLOWER_KEY_TCP_SRC_MASK 35
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_TCP_DST
-#define TCA_FLOWER_KEY_TCP_DST 19
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_TCP_DST_MASK
-#define TCA_FLOWER_KEY_TCP_DST_MASK 36
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_UDP_SRC
-#define TCA_FLOWER_KEY_UDP_SRC 20
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_UDP_SRC_MASK
-#define TCA_FLOWER_KEY_UDP_SRC_MASK 37
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_UDP_DST
-#define TCA_FLOWER_KEY_UDP_DST 21
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_UDP_DST_MASK
-#define TCA_FLOWER_KEY_UDP_DST_MASK 38
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_VLAN_ID
-#define TCA_FLOWER_KEY_VLAN_ID 23
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_VLAN_PRIO
-#define TCA_FLOWER_KEY_VLAN_PRIO 24
-#endif
-#ifndef HAVE_TCA_FLOWER_KEY_VLAN_ETH_TYPE
-#define TCA_FLOWER_KEY_VLAN_ETH_TYPE 25
-#endif
-
-/** Parser state definitions for mlx5_nl_flow_trans[]. */
-enum mlx5_nl_flow_trans {
-       INVALID,
-       BACK,
-       ATTR,
-       PATTERN,
-       ITEM_VOID,
-       ITEM_PORT_ID,
-       ITEM_ETH,
-       ITEM_VLAN,
-       ITEM_IPV4,
-       ITEM_IPV6,
-       ITEM_TCP,
-       ITEM_UDP,
-       ACTIONS,
-       ACTION_VOID,
-       ACTION_PORT_ID,
-       ACTION_DROP,
-       ACTION_OF_POP_VLAN,
-       ACTION_OF_PUSH_VLAN,
-       ACTION_OF_SET_VLAN_VID,
-       ACTION_OF_SET_VLAN_PCP,
-       END,
-};
-
-#define TRANS(...) (const enum mlx5_nl_flow_trans []){ __VA_ARGS__, INVALID, }
-
-#define PATTERN_COMMON \
-       ITEM_VOID, ITEM_PORT_ID, ACTIONS
-#define ACTIONS_COMMON \
-       ACTION_VOID, ACTION_OF_POP_VLAN, ACTION_OF_PUSH_VLAN, \
-       ACTION_OF_SET_VLAN_VID, ACTION_OF_SET_VLAN_PCP
-#define ACTIONS_FATE \
-       ACTION_PORT_ID, ACTION_DROP
-
-/** Parser state transitions used by mlx5_nl_flow_transpose(). */
-static const enum mlx5_nl_flow_trans *const mlx5_nl_flow_trans[] = {
-       [INVALID] = NULL,
-       [BACK] = NULL,
-       [ATTR] = TRANS(PATTERN),
-       [PATTERN] = TRANS(ITEM_ETH, PATTERN_COMMON),
-       [ITEM_VOID] = TRANS(BACK),
-       [ITEM_PORT_ID] = TRANS(BACK),
-       [ITEM_ETH] = TRANS(ITEM_IPV4, ITEM_IPV6, ITEM_VLAN, PATTERN_COMMON),
-       [ITEM_VLAN] = TRANS(ITEM_IPV4, ITEM_IPV6, PATTERN_COMMON),
-       [ITEM_IPV4] = TRANS(ITEM_TCP, ITEM_UDP, PATTERN_COMMON),
-       [ITEM_IPV6] = TRANS(ITEM_TCP, ITEM_UDP, PATTERN_COMMON),
-       [ITEM_TCP] = TRANS(PATTERN_COMMON),
-       [ITEM_UDP] = TRANS(PATTERN_COMMON),
-       [ACTIONS] = TRANS(ACTIONS_FATE, ACTIONS_COMMON),
-       [ACTION_VOID] = TRANS(BACK),
-       [ACTION_PORT_ID] = TRANS(ACTION_VOID, END),
-       [ACTION_DROP] = TRANS(ACTION_VOID, END),
-       [ACTION_OF_POP_VLAN] = TRANS(ACTIONS_FATE, ACTIONS_COMMON),
-       [ACTION_OF_PUSH_VLAN] = TRANS(ACTIONS_FATE, ACTIONS_COMMON),
-       [ACTION_OF_SET_VLAN_VID] = TRANS(ACTIONS_FATE, ACTIONS_COMMON),
-       [ACTION_OF_SET_VLAN_PCP] = TRANS(ACTIONS_FATE, ACTIONS_COMMON),
-       [END] = NULL,
-};
-
-/** Empty masks for known item types. */
-static const union {
-       struct rte_flow_item_port_id port_id;
-       struct rte_flow_item_eth eth;
-       struct rte_flow_item_vlan vlan;
-       struct rte_flow_item_ipv4 ipv4;
-       struct rte_flow_item_ipv6 ipv6;
-       struct rte_flow_item_tcp tcp;
-       struct rte_flow_item_udp udp;
-} mlx5_nl_flow_mask_empty;
-
-/** Supported masks for known item types. */
-static const struct {
-       struct rte_flow_item_port_id port_id;
-       struct rte_flow_item_eth eth;
-       struct rte_flow_item_vlan vlan;
-       struct rte_flow_item_ipv4 ipv4;
-       struct rte_flow_item_ipv6 ipv6;
-       struct rte_flow_item_tcp tcp;
-       struct rte_flow_item_udp udp;
-} mlx5_nl_flow_mask_supported = {
-       .port_id = {
-               .id = 0xffffffff,
-       },
-       .eth = {
-               .type = RTE_BE16(0xffff),
-               .dst.addr_bytes = "\xff\xff\xff\xff\xff\xff",
-               .src.addr_bytes = "\xff\xff\xff\xff\xff\xff",
-       },
-       .vlan = {
-               /* PCP and VID only, no DEI. */
-               .tci = RTE_BE16(0xefff),
-               .inner_type = RTE_BE16(0xffff),
-       },
-       .ipv4.hdr = {
-               .next_proto_id = 0xff,
-               .src_addr = RTE_BE32(0xffffffff),
-               .dst_addr = RTE_BE32(0xffffffff),
-       },
-       .ipv6.hdr = {
-               .proto = 0xff,
-               .src_addr =
-                       "\xff\xff\xff\xff\xff\xff\xff\xff"
-                       "\xff\xff\xff\xff\xff\xff\xff\xff",
-               .dst_addr =
-                       "\xff\xff\xff\xff\xff\xff\xff\xff"
-                       "\xff\xff\xff\xff\xff\xff\xff\xff",
-       },
-       .tcp.hdr = {
-               .src_port = RTE_BE16(0xffff),
-               .dst_port = RTE_BE16(0xffff),
-       },
-       .udp.hdr = {
-               .src_port = RTE_BE16(0xffff),
-               .dst_port = RTE_BE16(0xffff),
-       },
-};
-
-/**
- * Retrieve mask for pattern item.
- *
- * This function does basic sanity checks on a pattern item in order to
- * return the most appropriate mask for it.
- *
- * @param[in] item
- *   Item specification.
- * @param[in] mask_default
- *   Default mask for pattern item as specified by the flow API.
- * @param[in] mask_supported
- *   Mask fields supported by the implementation.
- * @param[in] mask_empty
- *   Empty mask to return when there is no specification.
- * @param[out] error
- *   Perform verbose error reporting if not NULL.
- *
- * @return
- *   Either @p item->mask or one of the mask parameters on success, NULL
- *   otherwise and rte_errno is set.
- */
-static const void *
-mlx5_nl_flow_item_mask(const struct rte_flow_item *item,
-                      const void *mask_default,
-                      const void *mask_supported,
-                      const void *mask_empty,
-                      size_t mask_size,
-                      struct rte_flow_error *error)
-{
-       const uint8_t *mask;
-       size_t i;
-
-       /* item->last and item->mask cannot exist without item->spec. */
-       if (!item->spec && (item->mask || item->last)) {
-               rte_flow_error_set
-                       (error, EINVAL, RTE_FLOW_ERROR_TYPE_ITEM, item,
-                        "\"mask\" or \"last\" field provided without a"
-                        " corresponding \"spec\"");
-               return NULL;
-       }
-       /* No spec, no mask, no problem. */
-       if (!item->spec)
-               return mask_empty;
-       mask = item->mask ? item->mask : mask_default;
-       assert(mask);
-       /*
-        * Single-pass check to make sure that:
-        * - Mask is supported, no bits are set outside mask_supported.
-        * - Both item->spec and item->last are included in mask.
-        */
-       for (i = 0; i != mask_size; ++i) {
-               if (!mask[i])
-                       continue;
-               if ((mask[i] | ((const uint8_t *)mask_supported)[i]) !=
-                   ((const uint8_t *)mask_supported)[i]) {
-                       rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask, "unsupported field found in \"mask\"");
-                       return NULL;
-               }
-               if (item->last &&
-                   (((const uint8_t *)item->spec)[i] & mask[i]) !=
-                   (((const uint8_t *)item->last)[i] & mask[i])) {
-                       rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_LAST,
-                                item->last,
-                                "range between \"spec\" and \"last\" not"
-                                " comprised in \"mask\"");
-                       return NULL;
-               }
-       }
-       return mask;
-}
-
-/**
- * Transpose flow rule description to rtnetlink message.
- *
- * This function transposes a flow rule description to a traffic control
- * (TC) filter creation message ready to be sent over Netlink.
- *
- * Target interface is specified as the first entry of the @p ptoi table.
- * Subsequent entries enable this function to resolve other DPDK port IDs
- * found in the flow rule.
- *
- * @param[out] buf
- *   Output message buffer. May be NULL when @p size is 0.
- * @param size
- *   Size of @p buf. Message may be truncated if not large enough.
- * @param[in] ptoi
- *   DPDK port ID to network interface index translation table. This table
- *   is terminated by an entry with a zero ifindex value.
- * @param[in] attr
- *   Flow rule attributes.
- * @param[in] pattern
- *   Pattern specification.
- * @param[in] actions
- *   Associated actions.
- * @param[out] error
- *   Perform verbose error reporting if not NULL.
- *
- * @return
- *   A positive value representing the exact size of the message in bytes
- *   regardless of the @p size parameter on success, a negative errno value
- *   otherwise and rte_errno is set.
- */
-int
-mlx5_nl_flow_transpose(void *buf,
-                      size_t size,
-                      const struct mlx5_nl_flow_ptoi *ptoi,
-                      const struct rte_flow_attr *attr,
-                      const struct rte_flow_item *pattern,
-                      const struct rte_flow_action *actions,
-                      struct rte_flow_error *error)
-{
-       alignas(struct nlmsghdr)
-       uint8_t buf_tmp[mnl_nlmsg_size(sizeof(struct tcmsg) + 1024)];
-       const struct rte_flow_item *item;
-       const struct rte_flow_action *action;
-       unsigned int n;
-       uint32_t act_index_cur;
-       bool in_port_id_set;
-       bool eth_type_set;
-       bool vlan_present;
-       bool vlan_eth_type_set;
-       bool ip_proto_set;
-       struct nlattr *na_flower;
-       struct nlattr *na_flower_act;
-       struct nlattr *na_vlan_id;
-       struct nlattr *na_vlan_priority;
-       const enum mlx5_nl_flow_trans *trans;
-       const enum mlx5_nl_flow_trans *back;
-
-       if (!size)
-               goto error_nobufs;
-init:
-       item = pattern;
-       action = actions;
-       n = 0;
-       act_index_cur = 0;
-       in_port_id_set = false;
-       eth_type_set = false;
-       vlan_present = false;
-       vlan_eth_type_set = false;
-       ip_proto_set = false;
-       na_flower = NULL;
-       na_flower_act = NULL;
-       na_vlan_id = NULL;
-       na_vlan_priority = NULL;
-       trans = TRANS(ATTR);
-       back = trans;
-trans:
-       switch (trans[n++]) {
-               union {
-                       const struct rte_flow_item_port_id *port_id;
-                       const struct rte_flow_item_eth *eth;
-                       const struct rte_flow_item_vlan *vlan;
-                       const struct rte_flow_item_ipv4 *ipv4;
-                       const struct rte_flow_item_ipv6 *ipv6;
-                       const struct rte_flow_item_tcp *tcp;
-                       const struct rte_flow_item_udp *udp;
-               } spec, mask;
-               union {
-                       const struct rte_flow_action_port_id *port_id;
-                       const struct rte_flow_action_of_push_vlan *of_push_vlan;
-                       const struct rte_flow_action_of_set_vlan_vid *
-                               of_set_vlan_vid;
-                       const struct rte_flow_action_of_set_vlan_pcp *
-                               of_set_vlan_pcp;
-               } conf;
-               struct nlmsghdr *nlh;
-               struct tcmsg *tcm;
-               struct nlattr *act_index;
-               struct nlattr *act;
-               unsigned int i;
-
-       case INVALID:
-               if (item->type)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM,
-                                item, "unsupported pattern item combination");
-               else if (action->type)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ACTION,
-                                action, "unsupported action combination");
-               return rte_flow_error_set
-                       (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
-                        "flow rule lacks some kind of fate action");
-       case BACK:
-               trans = back;
-               n = 0;
-               goto trans;
-       case ATTR:
-               /*
-                * Supported attributes: no groups, some priorities and
-                * ingress only. Don't care about transfer as it is the
-                * caller's problem.
-                */
-               if (attr->group)
-                       return rte_flow_error_set
-                               (error, ENOTSUP,
-                                RTE_FLOW_ERROR_TYPE_ATTR_GROUP,
-                                attr, "groups are not supported");
-               if (attr->priority > 0xfffe)
-                       return rte_flow_error_set
-                               (error, ENOTSUP,
-                                RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY,
-                                attr, "lowest priority level is 0xfffe");
-               if (!attr->ingress)
-                       return rte_flow_error_set
-                               (error, ENOTSUP,
-                                RTE_FLOW_ERROR_TYPE_ATTR_INGRESS,
-                                attr, "only ingress is supported");
-               if (attr->egress)
-                       return rte_flow_error_set
-                               (error, ENOTSUP,
-                                RTE_FLOW_ERROR_TYPE_ATTR_INGRESS,
-                                attr, "egress is not supported");
-               if (size < mnl_nlmsg_size(sizeof(*tcm)))
-                       goto error_nobufs;
-               nlh = mnl_nlmsg_put_header(buf);
-               nlh->nlmsg_type = 0;
-               nlh->nlmsg_flags = 0;
-               nlh->nlmsg_seq = 0;
-               tcm = mnl_nlmsg_put_extra_header(nlh, sizeof(*tcm));
-               tcm->tcm_family = AF_UNSPEC;
-               tcm->tcm_ifindex = ptoi[0].ifindex;
-               /*
-                * Let kernel pick a handle by default. A predictable handle
-                * can be set by the caller on the resulting buffer through
-                * mlx5_nl_flow_brand().
-                */
-               tcm->tcm_handle = 0;
-               tcm->tcm_parent = TC_H_MAKE(TC_H_INGRESS, TC_H_MIN_INGRESS);
-               /*
-                * Priority cannot be zero to prevent the kernel from
-                * picking one automatically.
-                */
-               tcm->tcm_info = TC_H_MAKE((attr->priority + 1) << 16,
-                                         RTE_BE16(ETH_P_ALL));
-               break;
-       case PATTERN:
-               if (!mnl_attr_put_strz_check(buf, size, TCA_KIND, "flower"))
-                       goto error_nobufs;
-               na_flower = mnl_attr_nest_start_check(buf, size, TCA_OPTIONS);
-               if (!na_flower)
-                       goto error_nobufs;
-               if (!mnl_attr_put_u32_check(buf, size, TCA_FLOWER_FLAGS,
-                                           TCA_CLS_FLAGS_SKIP_SW))
-                       goto error_nobufs;
-               break;
-       case ITEM_VOID:
-               if (item->type != RTE_FLOW_ITEM_TYPE_VOID)
-                       goto trans;
-               ++item;
-               break;
-       case ITEM_PORT_ID:
-               if (item->type != RTE_FLOW_ITEM_TYPE_PORT_ID)
-                       goto trans;
-               mask.port_id = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_port_id_mask,
-                        &mlx5_nl_flow_mask_supported.port_id,
-                        &mlx5_nl_flow_mask_empty.port_id,
-                        sizeof(mlx5_nl_flow_mask_supported.port_id), error);
-               if (!mask.port_id)
-                       return -rte_errno;
-               if (mask.port_id == &mlx5_nl_flow_mask_empty.port_id) {
-                       in_port_id_set = 1;
-                       ++item;
-                       break;
-               }
-               spec.port_id = item->spec;
-               if (mask.port_id->id && mask.port_id->id != 0xffffffff)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask.port_id,
-                                "no support for partial mask on"
-                                " \"id\" field");
-               if (!mask.port_id->id)
-                       i = 0;
-               else
-                       for (i = 0; ptoi[i].ifindex; ++i)
-                               if (ptoi[i].port_id == spec.port_id->id)
-                                       break;
-               if (!ptoi[i].ifindex)
-                       return rte_flow_error_set
-                               (error, ENODEV, RTE_FLOW_ERROR_TYPE_ITEM_SPEC,
-                                spec.port_id,
-                                "missing data to convert port ID to ifindex");
-               tcm = mnl_nlmsg_get_payload(buf);
-               if (in_port_id_set &&
-                   ptoi[i].ifindex != (unsigned int)tcm->tcm_ifindex)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_SPEC,
-                                spec.port_id,
-                                "cannot match traffic for several port IDs"
-                                " through a single flow rule");
-               tcm->tcm_ifindex = ptoi[i].ifindex;
-               in_port_id_set = 1;
-               ++item;
-               break;
-       case ITEM_ETH:
-               if (item->type != RTE_FLOW_ITEM_TYPE_ETH)
-                       goto trans;
-               mask.eth = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_eth_mask,
-                        &mlx5_nl_flow_mask_supported.eth,
-                        &mlx5_nl_flow_mask_empty.eth,
-                        sizeof(mlx5_nl_flow_mask_supported.eth), error);
-               if (!mask.eth)
-                       return -rte_errno;
-               if (mask.eth == &mlx5_nl_flow_mask_empty.eth) {
-                       ++item;
-                       break;
-               }
-               spec.eth = item->spec;
-               if (mask.eth->type && mask.eth->type != RTE_BE16(0xffff))
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask.eth,
-                                "no support for partial mask on"
-                                " \"type\" field");
-               if (mask.eth->type) {
-                       if (!mnl_attr_put_u16_check(buf, size,
-                                                   TCA_FLOWER_KEY_ETH_TYPE,
-                                                   spec.eth->type))
-                               goto error_nobufs;
-                       eth_type_set = 1;
-               }
-               if ((!is_zero_ether_addr(&mask.eth->dst) &&
-                    (!mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_ETH_DST,
-                                         ETHER_ADDR_LEN,
-                                         spec.eth->dst.addr_bytes) ||
-                     !mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_ETH_DST_MASK,
-                                         ETHER_ADDR_LEN,
-                                         mask.eth->dst.addr_bytes))) ||
-                   (!is_zero_ether_addr(&mask.eth->src) &&
-                    (!mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_ETH_SRC,
-                                         ETHER_ADDR_LEN,
-                                         spec.eth->src.addr_bytes) ||
-                     !mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_ETH_SRC_MASK,
-                                         ETHER_ADDR_LEN,
-                                         mask.eth->src.addr_bytes))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ITEM_VLAN:
-               if (item->type != RTE_FLOW_ITEM_TYPE_VLAN)
-                       goto trans;
-               mask.vlan = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_vlan_mask,
-                        &mlx5_nl_flow_mask_supported.vlan,
-                        &mlx5_nl_flow_mask_empty.vlan,
-                        sizeof(mlx5_nl_flow_mask_supported.vlan), error);
-               if (!mask.vlan)
-                       return -rte_errno;
-               if (!eth_type_set &&
-                   !mnl_attr_put_u16_check(buf, size,
-                                           TCA_FLOWER_KEY_ETH_TYPE,
-                                           RTE_BE16(ETH_P_8021Q)))
-                       goto error_nobufs;
-               eth_type_set = 1;
-               vlan_present = 1;
-               if (mask.vlan == &mlx5_nl_flow_mask_empty.vlan) {
-                       ++item;
-                       break;
-               }
-               spec.vlan = item->spec;
-               if ((mask.vlan->tci & RTE_BE16(0xe000) &&
-                    (mask.vlan->tci & RTE_BE16(0xe000)) != RTE_BE16(0xe000)) ||
-                   (mask.vlan->tci & RTE_BE16(0x0fff) &&
-                    (mask.vlan->tci & RTE_BE16(0x0fff)) != RTE_BE16(0x0fff)) ||
-                   (mask.vlan->inner_type &&
-                    mask.vlan->inner_type != RTE_BE16(0xffff)))
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask.vlan,
-                                "no support for partial masks on"
-                                " \"tci\" (PCP and VID parts) and"
-                                " \"inner_type\" fields");
-               if (mask.vlan->inner_type) {
-                       if (!mnl_attr_put_u16_check
-                           (buf, size, TCA_FLOWER_KEY_VLAN_ETH_TYPE,
-                            spec.vlan->inner_type))
-                               goto error_nobufs;
-                       vlan_eth_type_set = 1;
-               }
-               if ((mask.vlan->tci & RTE_BE16(0xe000) &&
-                    !mnl_attr_put_u8_check
-                    (buf, size, TCA_FLOWER_KEY_VLAN_PRIO,
-                     (rte_be_to_cpu_16(spec.vlan->tci) >> 13) & 0x7)) ||
-                   (mask.vlan->tci & RTE_BE16(0x0fff) &&
-                    !mnl_attr_put_u16_check
-                    (buf, size, TCA_FLOWER_KEY_VLAN_ID,
-                     rte_be_to_cpu_16(spec.vlan->tci & RTE_BE16(0x0fff)))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ITEM_IPV4:
-               if (item->type != RTE_FLOW_ITEM_TYPE_IPV4)
-                       goto trans;
-               mask.ipv4 = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_ipv4_mask,
-                        &mlx5_nl_flow_mask_supported.ipv4,
-                        &mlx5_nl_flow_mask_empty.ipv4,
-                        sizeof(mlx5_nl_flow_mask_supported.ipv4), error);
-               if (!mask.ipv4)
-                       return -rte_errno;
-               if ((!eth_type_set || !vlan_eth_type_set) &&
-                   !mnl_attr_put_u16_check(buf, size,
-                                           vlan_present ?
-                                           TCA_FLOWER_KEY_VLAN_ETH_TYPE :
-                                           TCA_FLOWER_KEY_ETH_TYPE,
-                                           RTE_BE16(ETH_P_IP)))
-                       goto error_nobufs;
-               eth_type_set = 1;
-               vlan_eth_type_set = 1;
-               if (mask.ipv4 == &mlx5_nl_flow_mask_empty.ipv4) {
-                       ++item;
-                       break;
-               }
-               spec.ipv4 = item->spec;
-               if (mask.ipv4->hdr.next_proto_id &&
-                   mask.ipv4->hdr.next_proto_id != 0xff)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask.ipv4,
-                                "no support for partial mask on"
-                                " \"hdr.next_proto_id\" field");
-               if (mask.ipv4->hdr.next_proto_id) {
-                       if (!mnl_attr_put_u8_check
-                           (buf, size, TCA_FLOWER_KEY_IP_PROTO,
-                            spec.ipv4->hdr.next_proto_id))
-                               goto error_nobufs;
-                       ip_proto_set = 1;
-               }
-               if ((mask.ipv4->hdr.src_addr &&
-                    (!mnl_attr_put_u32_check(buf, size,
-                                             TCA_FLOWER_KEY_IPV4_SRC,
-                                             spec.ipv4->hdr.src_addr) ||
-                     !mnl_attr_put_u32_check(buf, size,
-                                             TCA_FLOWER_KEY_IPV4_SRC_MASK,
-                                             mask.ipv4->hdr.src_addr))) ||
-                   (mask.ipv4->hdr.dst_addr &&
-                    (!mnl_attr_put_u32_check(buf, size,
-                                             TCA_FLOWER_KEY_IPV4_DST,
-                                             spec.ipv4->hdr.dst_addr) ||
-                     !mnl_attr_put_u32_check(buf, size,
-                                             TCA_FLOWER_KEY_IPV4_DST_MASK,
-                                             mask.ipv4->hdr.dst_addr))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ITEM_IPV6:
-               if (item->type != RTE_FLOW_ITEM_TYPE_IPV6)
-                       goto trans;
-               mask.ipv6 = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_ipv6_mask,
-                        &mlx5_nl_flow_mask_supported.ipv6,
-                        &mlx5_nl_flow_mask_empty.ipv6,
-                        sizeof(mlx5_nl_flow_mask_supported.ipv6), error);
-               if (!mask.ipv6)
-                       return -rte_errno;
-               if ((!eth_type_set || !vlan_eth_type_set) &&
-                   !mnl_attr_put_u16_check(buf, size,
-                                           vlan_present ?
-                                           TCA_FLOWER_KEY_VLAN_ETH_TYPE :
-                                           TCA_FLOWER_KEY_ETH_TYPE,
-                                           RTE_BE16(ETH_P_IPV6)))
-                       goto error_nobufs;
-               eth_type_set = 1;
-               vlan_eth_type_set = 1;
-               if (mask.ipv6 == &mlx5_nl_flow_mask_empty.ipv6) {
-                       ++item;
-                       break;
-               }
-               spec.ipv6 = item->spec;
-               if (mask.ipv6->hdr.proto && mask.ipv6->hdr.proto != 0xff)
-                       return rte_flow_error_set
-                               (error, ENOTSUP, RTE_FLOW_ERROR_TYPE_ITEM_MASK,
-                                mask.ipv6,
-                                "no support for partial mask on"
-                                " \"hdr.proto\" field");
-               if (mask.ipv6->hdr.proto) {
-                       if (!mnl_attr_put_u8_check
-                           (buf, size, TCA_FLOWER_KEY_IP_PROTO,
-                            spec.ipv6->hdr.proto))
-                               goto error_nobufs;
-                       ip_proto_set = 1;
-               }
-               if ((!IN6_IS_ADDR_UNSPECIFIED(mask.ipv6->hdr.src_addr) &&
-                    (!mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_IPV6_SRC,
-                                         sizeof(spec.ipv6->hdr.src_addr),
-                                         spec.ipv6->hdr.src_addr) ||
-                     !mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_IPV6_SRC_MASK,
-                                         sizeof(mask.ipv6->hdr.src_addr),
-                                         mask.ipv6->hdr.src_addr))) ||
-                   (!IN6_IS_ADDR_UNSPECIFIED(mask.ipv6->hdr.dst_addr) &&
-                    (!mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_IPV6_DST,
-                                         sizeof(spec.ipv6->hdr.dst_addr),
-                                         spec.ipv6->hdr.dst_addr) ||
-                     !mnl_attr_put_check(buf, size,
-                                         TCA_FLOWER_KEY_IPV6_DST_MASK,
-                                         sizeof(mask.ipv6->hdr.dst_addr),
-                                         mask.ipv6->hdr.dst_addr))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ITEM_TCP:
-               if (item->type != RTE_FLOW_ITEM_TYPE_TCP)
-                       goto trans;
-               mask.tcp = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_tcp_mask,
-                        &mlx5_nl_flow_mask_supported.tcp,
-                        &mlx5_nl_flow_mask_empty.tcp,
-                        sizeof(mlx5_nl_flow_mask_supported.tcp), error);
-               if (!mask.tcp)
-                       return -rte_errno;
-               if (!ip_proto_set &&
-                   !mnl_attr_put_u8_check(buf, size,
-                                          TCA_FLOWER_KEY_IP_PROTO,
-                                          IPPROTO_TCP))
-                       goto error_nobufs;
-               if (mask.tcp == &mlx5_nl_flow_mask_empty.tcp) {
-                       ++item;
-                       break;
-               }
-               spec.tcp = item->spec;
-               if ((mask.tcp->hdr.src_port &&
-                    (!mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_TCP_SRC,
-                                             spec.tcp->hdr.src_port) ||
-                     !mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_TCP_SRC_MASK,
-                                             mask.tcp->hdr.src_port))) ||
-                   (mask.tcp->hdr.dst_port &&
-                    (!mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_TCP_DST,
-                                             spec.tcp->hdr.dst_port) ||
-                     !mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_TCP_DST_MASK,
-                                             mask.tcp->hdr.dst_port))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ITEM_UDP:
-               if (item->type != RTE_FLOW_ITEM_TYPE_UDP)
-                       goto trans;
-               mask.udp = mlx5_nl_flow_item_mask
-                       (item, &rte_flow_item_udp_mask,
-                        &mlx5_nl_flow_mask_supported.udp,
-                        &mlx5_nl_flow_mask_empty.udp,
-                        sizeof(mlx5_nl_flow_mask_supported.udp), error);
-               if (!mask.udp)
-                       return -rte_errno;
-               if (!ip_proto_set &&
-                   !mnl_attr_put_u8_check(buf, size,
-                                          TCA_FLOWER_KEY_IP_PROTO,
-                                          IPPROTO_UDP))
-                       goto error_nobufs;
-               if (mask.udp == &mlx5_nl_flow_mask_empty.udp) {
-                       ++item;
-                       break;
-               }
-               spec.udp = item->spec;
-               if ((mask.udp->hdr.src_port &&
-                    (!mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_UDP_SRC,
-                                             spec.udp->hdr.src_port) ||
-                     !mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_UDP_SRC_MASK,
-                                             mask.udp->hdr.src_port))) ||
-                   (mask.udp->hdr.dst_port &&
-                    (!mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_UDP_DST,
-                                             spec.udp->hdr.dst_port) ||
-                     !mnl_attr_put_u16_check(buf, size,
-                                             TCA_FLOWER_KEY_UDP_DST_MASK,
-                                             mask.udp->hdr.dst_port))))
-                       goto error_nobufs;
-               ++item;
-               break;
-       case ACTIONS:
-               if (item->type != RTE_FLOW_ITEM_TYPE_END)
-                       goto trans;
-               assert(na_flower);
-               assert(!na_flower_act);
-               na_flower_act =
-                       mnl_attr_nest_start_check(buf, size, TCA_FLOWER_ACT);
-               if (!na_flower_act)
-                       goto error_nobufs;
-               act_index_cur = 1;
-               break;
-       case ACTION_VOID:
-               if (action->type != RTE_FLOW_ACTION_TYPE_VOID)
-                       goto trans;
-               ++action;
-               break;
-       case ACTION_PORT_ID:
-               if (action->type != RTE_FLOW_ACTION_TYPE_PORT_ID)
-                       goto trans;
-               conf.port_id = action->conf;
-               if (conf.port_id->original)
-                       i = 0;
-               else
-                       for (i = 0; ptoi[i].ifindex; ++i)
-                               if (ptoi[i].port_id == conf.port_id->id)
-                                       break;
-               if (!ptoi[i].ifindex)
-                       return rte_flow_error_set
-                               (error, ENODEV, RTE_FLOW_ERROR_TYPE_ACTION_CONF,
-                                conf.port_id,
-                                "missing data to convert port ID to ifindex");
-               act_index =
-                       mnl_attr_nest_start_check(buf, size, act_index_cur++);
-               if (!act_index ||
-                   !mnl_attr_put_strz_check(buf, size, TCA_ACT_KIND, "mirred"))
-                       goto error_nobufs;
-               act = mnl_attr_nest_start_check(buf, size, TCA_ACT_OPTIONS);
-               if (!act)
-                       goto error_nobufs;
-               if (!mnl_attr_put_check(buf, size, TCA_MIRRED_PARMS,
-                                       sizeof(struct tc_mirred),
-                                       &(struct tc_mirred){
-                                               .action = TC_ACT_STOLEN,
-                                               .eaction = TCA_EGRESS_REDIR,
-                                               .ifindex = ptoi[i].ifindex,
-                                       }))
-                       goto error_nobufs;
-               mnl_attr_nest_end(buf, act);
-               mnl_attr_nest_end(buf, act_index);
-               ++action;
-               break;
-       case ACTION_DROP:
-               if (action->type != RTE_FLOW_ACTION_TYPE_DROP)
-                       goto trans;
-               act_index =
-                       mnl_attr_nest_start_check(buf, size, act_index_cur++);
-               if (!act_index ||
-                   !mnl_attr_put_strz_check(buf, size, TCA_ACT_KIND, "gact"))
-                       goto error_nobufs;
-               act = mnl_attr_nest_start_check(buf, size, TCA_ACT_OPTIONS);
-               if (!act)
-                       goto error_nobufs;
-               if (!mnl_attr_put_check(buf, size, TCA_GACT_PARMS,
-                                       sizeof(struct tc_gact),
-                                       &(struct tc_gact){
-                                               .action = TC_ACT_SHOT,
-                                       }))
-                       goto error_nobufs;
-               mnl_attr_nest_end(buf, act);
-               mnl_attr_nest_end(buf, act_index);
-               ++action;
-               break;
-       case ACTION_OF_POP_VLAN:
-               if (action->type != RTE_FLOW_ACTION_TYPE_OF_POP_VLAN)
-                       goto trans;
-               conf.of_push_vlan = NULL;
-               i = TCA_VLAN_ACT_POP;
-               goto action_of_vlan;
-       case ACTION_OF_PUSH_VLAN:
-               if (action->type != RTE_FLOW_ACTION_TYPE_OF_PUSH_VLAN)
-                       goto trans;
-               conf.of_push_vlan = action->conf;
-               i = TCA_VLAN_ACT_PUSH;
-               goto action_of_vlan;
-       case ACTION_OF_SET_VLAN_VID:
-               if (action->type != RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID)
-                       goto trans;
-               conf.of_set_vlan_vid = action->conf;
-               if (na_vlan_id)
-                       goto override_na_vlan_id;
-               i = TCA_VLAN_ACT_MODIFY;
-               goto action_of_vlan;
-       case ACTION_OF_SET_VLAN_PCP:
-               if (action->type != RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP)
-                       goto trans;
-               conf.of_set_vlan_pcp = action->conf;
-               if (na_vlan_priority)
-                       goto override_na_vlan_priority;
-               i = TCA_VLAN_ACT_MODIFY;
-               goto action_of_vlan;
-action_of_vlan:
-               act_index =
-                       mnl_attr_nest_start_check(buf, size, act_index_cur++);
-               if (!act_index ||
-                   !mnl_attr_put_strz_check(buf, size, TCA_ACT_KIND, "vlan"))
-                       goto error_nobufs;
-               act = mnl_attr_nest_start_check(buf, size, TCA_ACT_OPTIONS);
-               if (!act)
-                       goto error_nobufs;
-               if (!mnl_attr_put_check(buf, size, TCA_VLAN_PARMS,
-                                       sizeof(struct tc_vlan),
-                                       &(struct tc_vlan){
-                                               .action = TC_ACT_PIPE,
-                                               .v_action = i,
-                                       }))
-                       goto error_nobufs;
-               if (i == TCA_VLAN_ACT_POP) {
-                       mnl_attr_nest_end(buf, act);
-                       mnl_attr_nest_end(buf, act_index);
-                       ++action;
-                       break;
-               }
-               if (i == TCA_VLAN_ACT_PUSH &&
-                   !mnl_attr_put_u16_check(buf, size,
-                                           TCA_VLAN_PUSH_VLAN_PROTOCOL,
-                                           conf.of_push_vlan->ethertype))
-                       goto error_nobufs;
-               na_vlan_id = mnl_nlmsg_get_payload_tail(buf);
-               if (!mnl_attr_put_u16_check(buf, size, TCA_VLAN_PAD, 0))
-                       goto error_nobufs;
-               na_vlan_priority = mnl_nlmsg_get_payload_tail(buf);
-               if (!mnl_attr_put_u8_check(buf, size, TCA_VLAN_PAD, 0))
-                       goto error_nobufs;
-               mnl_attr_nest_end(buf, act);
-               mnl_attr_nest_end(buf, act_index);
-               if (action->type == RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_VID) {
-override_na_vlan_id:
-                       na_vlan_id->nla_type = TCA_VLAN_PUSH_VLAN_ID;
-                       *(uint16_t *)mnl_attr_get_payload(na_vlan_id) =
-                               rte_be_to_cpu_16
-                               (conf.of_set_vlan_vid->vlan_vid);
-               } else if (action->type ==
-                          RTE_FLOW_ACTION_TYPE_OF_SET_VLAN_PCP) {
-override_na_vlan_priority:
-                       na_vlan_priority->nla_type =
-                               TCA_VLAN_PUSH_VLAN_PRIORITY;
-                       *(uint8_t *)mnl_attr_get_payload(na_vlan_priority) =
-                               conf.of_set_vlan_pcp->vlan_pcp;
-               }
-               ++action;
-               break;
-       case END:
-               if (item->type != RTE_FLOW_ITEM_TYPE_END ||
-                   action->type != RTE_FLOW_ACTION_TYPE_END)
-                       goto trans;
-               if (na_flower_act)
-                       mnl_attr_nest_end(buf, na_flower_act);
-               if (na_flower)
-                       mnl_attr_nest_end(buf, na_flower);
-               nlh = buf;
-               return nlh->nlmsg_len;
-       }
-       back = trans;
-       trans = mlx5_nl_flow_trans[trans[n - 1]];
-       n = 0;
-       goto trans;
-error_nobufs:
-       if (buf != buf_tmp) {
-               buf = buf_tmp;
-               size = sizeof(buf_tmp);
-               goto init;
-       }
-       return rte_flow_error_set
-               (error, ENOBUFS, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
-                "generated TC message is too large");
-}
-
-/**
- * Brand rtnetlink buffer with unique handle.
- *
- * This handle should be unique for a given network interface to avoid
- * collisions.
- *
- * @param buf
- *   Flow rule buffer previously initialized by mlx5_nl_flow_transpose().
- * @param handle
- *   Unique 32-bit handle to use.
- */
-void
-mlx5_nl_flow_brand(void *buf, uint32_t handle)
-{
-       struct tcmsg *tcm = mnl_nlmsg_get_payload(buf);
-
-       tcm->tcm_handle = handle;
-}
-
-/**
- * Send Netlink message with acknowledgment.
- *
- * @param nl
- *   Libmnl socket to use.
- * @param nlh
- *   Message to send. This function always raises the NLM_F_ACK flag before
- *   sending.
- *
- * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
- */
-static int
-mlx5_nl_flow_nl_ack(struct mnl_socket *nl, struct nlmsghdr *nlh)
-{
-       alignas(struct nlmsghdr)
-       uint8_t ans[mnl_nlmsg_size(sizeof(struct nlmsgerr)) +
-                   nlh->nlmsg_len - sizeof(*nlh)];
-       uint32_t seq = random();
-       int ret;
-
-       nlh->nlmsg_flags |= NLM_F_ACK;
-       nlh->nlmsg_seq = seq;
-       ret = mnl_socket_sendto(nl, nlh, nlh->nlmsg_len);
-       if (ret != -1)
-               ret = mnl_socket_recvfrom(nl, ans, sizeof(ans));
-       if (ret != -1)
-               ret = mnl_cb_run
-                       (ans, ret, seq, mnl_socket_get_portid(nl), NULL, NULL);
-       if (!ret)
-               return 0;
-       rte_errno = errno;
-       return -rte_errno;
-}
-
-/**
- * Create a Netlink flow rule.
- *
- * @param nl
- *   Libmnl socket to use.
- * @param buf
- *   Flow rule buffer previously initialized by mlx5_nl_flow_transpose().
- * @param[out] error
- *   Perform verbose error reporting if not NULL.
- *
- * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
- */
-int
-mlx5_nl_flow_create(struct mnl_socket *nl, void *buf,
-                   struct rte_flow_error *error)
-{
-       struct nlmsghdr *nlh = buf;
-
-       nlh->nlmsg_type = RTM_NEWTFILTER;
-       nlh->nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL;
-       if (!mlx5_nl_flow_nl_ack(nl, nlh))
-               return 0;
-       return rte_flow_error_set
-               (error, rte_errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
-                "netlink: failed to create TC flow rule");
-}
-
-/**
- * Destroy a Netlink flow rule.
- *
- * @param nl
- *   Libmnl socket to use.
- * @param buf
- *   Flow rule buffer previously initialized by mlx5_nl_flow_transpose().
- * @param[out] error
- *   Perform verbose error reporting if not NULL.
- *
- * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
- */
-int
-mlx5_nl_flow_destroy(struct mnl_socket *nl, void *buf,
-                    struct rte_flow_error *error)
-{
-       struct nlmsghdr *nlh = buf;
-
-       nlh->nlmsg_type = RTM_DELTFILTER;
-       nlh->nlmsg_flags = NLM_F_REQUEST;
-       if (!mlx5_nl_flow_nl_ack(nl, nlh))
-               return 0;
-       return rte_flow_error_set
-               (error, errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED, NULL,
-                "netlink: failed to destroy TC flow rule");
-}
-
-/**
- * Initialize ingress qdisc of a given network interface.
- *
- * @param nl
- *   Libmnl socket of the @p NETLINK_ROUTE kind.
- * @param ifindex
- *   Index of network interface to initialize.
- * @param[out] error
- *   Perform verbose error reporting if not NULL.
- *
- * @return
- *   0 on success, a negative errno value otherwise and rte_errno is set.
- */
-int
-mlx5_nl_flow_init(struct mnl_socket *nl, unsigned int ifindex,
-                 struct rte_flow_error *error)
-{
-       struct nlmsghdr *nlh;
-       struct tcmsg *tcm;
-       alignas(struct nlmsghdr)
-       uint8_t buf[mnl_nlmsg_size(sizeof(*tcm) + 128)];
-
-       /* Destroy existing ingress qdisc and everything attached to it. */
-       nlh = mnl_nlmsg_put_header(buf);
-       nlh->nlmsg_type = RTM_DELQDISC;
-       nlh->nlmsg_flags = NLM_F_REQUEST;
-       tcm = mnl_nlmsg_put_extra_header(nlh, sizeof(*tcm));
-       tcm->tcm_family = AF_UNSPEC;
-       tcm->tcm_ifindex = ifindex;
-       tcm->tcm_handle = TC_H_MAKE(TC_H_INGRESS, 0);
-       tcm->tcm_parent = TC_H_INGRESS;
-       /* Ignore errors when qdisc is already absent. */
-       if (mlx5_nl_flow_nl_ack(nl, nlh) &&
-           rte_errno != EINVAL && rte_errno != ENOENT)
-               return rte_flow_error_set
-                       (error, rte_errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-                        NULL, "netlink: failed to remove ingress qdisc");
-       /* Create fresh ingress qdisc. */
-       nlh = mnl_nlmsg_put_header(buf);
-       nlh->nlmsg_type = RTM_NEWQDISC;
-       nlh->nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL;
-       tcm = mnl_nlmsg_put_extra_header(nlh, sizeof(*tcm));
-       tcm->tcm_family = AF_UNSPEC;
-       tcm->tcm_ifindex = ifindex;
-       tcm->tcm_handle = TC_H_MAKE(TC_H_INGRESS, 0);
-       tcm->tcm_parent = TC_H_INGRESS;
-       mnl_attr_put_strz_check(nlh, sizeof(buf), TCA_KIND, "ingress");
-       if (mlx5_nl_flow_nl_ack(nl, nlh))
-               return rte_flow_error_set
-                       (error, rte_errno, RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
-                        NULL, "netlink: failed to create ingress qdisc");
-       return 0;
-}
-
-/**
- * Create and configure a libmnl socket for Netlink flow rules.
- *
- * @return
- *   A valid libmnl socket object pointer on success, NULL otherwise and
- *   rte_errno is set.
- */
-struct mnl_socket *
-mlx5_nl_flow_socket_create(void)
-{
-       struct mnl_socket *nl = mnl_socket_open(NETLINK_ROUTE);
-
-       if (nl) {
-               mnl_socket_setsockopt(nl, NETLINK_CAP_ACK, &(int){ 1 },
-                                     sizeof(int));
-               if (!mnl_socket_bind(nl, 0, MNL_SOCKET_AUTOPID))
-                       return nl;
-       }
-       rte_errno = errno;
-       if (nl)
-               mnl_socket_close(nl);
-       return NULL;
-}
-
-/**
- * Destroy a libmnl socket.
- */
-void
-mlx5_nl_flow_socket_destroy(struct mnl_socket *nl)
-{
-       mnl_socket_close(nl);
-}