net: add rte prefix to ether structures
[dpdk.git] / lib / librte_pipeline / rte_table_action.c
index 8bcc4eb..1e85cc9 100644 (file)
@@ -1,7 +1,6 @@
 /* SPDX-License-Identifier: BSD-3-Clause
  * Copyright(c) 2010-2018 Intel Corporation
  */
-
 #include <stdlib.h>
 #include <string.h>
 
@@ -15,6 +14,8 @@
 #include <rte_esp.h>
 #include <rte_tcp.h>
 #include <rte_udp.h>
+#include <rte_cryptodev.h>
+#include <rte_cryptodev_pmd.h>
 
 #include "rte_table_action.h"
 
@@ -44,6 +45,55 @@ fwd_apply(struct fwd_data *data,
        return 0;
 }
 
+/**
+ * RTE_TABLE_ACTION_LB
+ */
+static int
+lb_cfg_check(struct rte_table_action_lb_config *cfg)
+{
+       if ((cfg == NULL) ||
+               (cfg->key_size < RTE_TABLE_ACTION_LB_KEY_SIZE_MIN) ||
+               (cfg->key_size > RTE_TABLE_ACTION_LB_KEY_SIZE_MAX) ||
+               (!rte_is_power_of_2(cfg->key_size)) ||
+               (cfg->f_hash == NULL))
+               return -1;
+
+       return 0;
+}
+
+struct lb_data {
+       uint32_t out[RTE_TABLE_ACTION_LB_TABLE_SIZE];
+} __attribute__((__packed__));
+
+static int
+lb_apply(struct lb_data *data,
+       struct rte_table_action_lb_params *p)
+{
+       memcpy(data->out, p->out, sizeof(data->out));
+
+       return 0;
+}
+
+static __rte_always_inline void
+pkt_work_lb(struct rte_mbuf *mbuf,
+       struct lb_data *data,
+       struct rte_table_action_lb_config *cfg)
+{
+       uint8_t *pkt_key = RTE_MBUF_METADATA_UINT8_PTR(mbuf, cfg->key_offset);
+       uint32_t *out = RTE_MBUF_METADATA_UINT32_PTR(mbuf, cfg->out_offset);
+       uint64_t digest, pos;
+       uint32_t out_val;
+
+       digest = cfg->f_hash(pkt_key,
+               cfg->key_mask,
+               cfg->key_size,
+               cfg->seed);
+       pos = digest & (RTE_TABLE_ACTION_LB_TABLE_SIZE - 1);
+       out_val = data->out[pos];
+
+       *out = out_val;
+}
+
 /**
  * RTE_TABLE_ACTION_MTR
  */
@@ -57,39 +107,31 @@ mtr_cfg_check(struct rte_table_action_mtr_config *mtr)
        return 0;
 }
 
-#define MBUF_SCHED_QUEUE_TC_COLOR(queue, tc, color)        \
-       ((uint16_t)((((uint64_t)(queue)) & 0x3) |          \
-       ((((uint64_t)(tc)) & 0x3) << 2) |                  \
-       ((((uint64_t)(color)) & 0x3) << 4)))
-
-#define MBUF_SCHED_COLOR(sched, color)                     \
-       (((sched) & (~0x30LLU)) | ((color) << 4))
-
 struct mtr_trtcm_data {
        struct rte_meter_trtcm trtcm;
-       uint64_t stats[e_RTE_METER_COLORS];
+       uint64_t stats[RTE_COLORS];
 } __attribute__((__packed__));
 
 #define MTR_TRTCM_DATA_METER_PROFILE_ID_GET(data)          \
-       (((data)->stats[e_RTE_METER_GREEN] & 0xF8LLU) >> 3)
+       (((data)->stats[RTE_COLOR_GREEN] & 0xF8LLU) >> 3)
 
 static void
 mtr_trtcm_data_meter_profile_id_set(struct mtr_trtcm_data *data,
        uint32_t profile_id)
 {
-       data->stats[e_RTE_METER_GREEN] &= ~0xF8LLU;
-       data->stats[e_RTE_METER_GREEN] |= (profile_id % 32) << 3;
+       data->stats[RTE_COLOR_GREEN] &= ~0xF8LLU;
+       data->stats[RTE_COLOR_GREEN] |= (profile_id % 32) << 3;
 }
 
 #define MTR_TRTCM_DATA_POLICER_ACTION_DROP_GET(data, color)\
        (((data)->stats[(color)] & 4LLU) >> 2)
 
 #define MTR_TRTCM_DATA_POLICER_ACTION_COLOR_GET(data, color)\
-       ((enum rte_meter_color)((data)->stats[(color)] & 3LLU))
+       ((enum rte_color)((data)->stats[(color)] & 3LLU))
 
 static void
 mtr_trtcm_data_policer_action_set(struct mtr_trtcm_data *data,
-       enum rte_meter_color color,
+       enum rte_color color,
        enum rte_table_action_policer action)
 {
        if (action == RTE_TABLE_ACTION_POLICER_DROP) {
@@ -102,14 +144,14 @@ mtr_trtcm_data_policer_action_set(struct mtr_trtcm_data *data,
 
 static uint64_t
 mtr_trtcm_data_stats_get(struct mtr_trtcm_data *data,
-       enum rte_meter_color color)
+       enum rte_color color)
 {
        return data->stats[color] >> 8;
 }
 
 static void
 mtr_trtcm_data_stats_reset(struct mtr_trtcm_data *data,
-       enum rte_meter_color color)
+       enum rte_color color)
 {
        data->stats[color] &= 0xFFLU;
 }
@@ -124,9 +166,9 @@ mtr_data_size(struct rte_table_action_mtr_config *mtr)
 }
 
 struct dscp_table_entry_data {
-       enum rte_meter_color color;
+       enum rte_color color;
        uint16_t tc;
-       uint16_t queue_tc_color;
+       uint16_t tc_queue;
 };
 
 struct dscp_table_data {
@@ -245,16 +287,16 @@ mtr_apply(struct mtr_trtcm_data *data,
 
                /* Policer actions */
                mtr_trtcm_data_policer_action_set(data_tc,
-                       e_RTE_METER_GREEN,
-                       p_tc->policer[e_RTE_METER_GREEN]);
+                       RTE_COLOR_GREEN,
+                       p_tc->policer[RTE_COLOR_GREEN]);
 
                mtr_trtcm_data_policer_action_set(data_tc,
-                       e_RTE_METER_YELLOW,
-                       p_tc->policer[e_RTE_METER_YELLOW]);
+                       RTE_COLOR_YELLOW,
+                       p_tc->policer[RTE_COLOR_YELLOW]);
 
                mtr_trtcm_data_policer_action_set(data_tc,
-                       e_RTE_METER_RED,
-                       p_tc->policer[e_RTE_METER_RED]);
+                       RTE_COLOR_RED,
+                       p_tc->policer[RTE_COLOR_RED]);
        }
 
        return 0;
@@ -269,17 +311,15 @@ pkt_work_mtr(struct rte_mbuf *mbuf,
        uint32_t dscp,
        uint16_t total_length)
 {
-       uint64_t drop_mask, sched;
-       uint64_t *sched_ptr = (uint64_t *) &mbuf->hash.sched;
+       uint64_t drop_mask;
        struct dscp_table_entry_data *dscp_entry = &dscp_table->entry[dscp];
-       enum rte_meter_color color_in, color_meter, color_policer;
+       enum rte_color color_in, color_meter, color_policer;
        uint32_t tc, mp_id;
 
        tc = dscp_entry->tc;
        color_in = dscp_entry->color;
        data += tc;
        mp_id = MTR_TRTCM_DATA_METER_PROFILE_ID_GET(data);
-       sched = *sched_ptr;
 
        /* Meter */
        color_meter = rte_meter_trtcm_color_aware_check(
@@ -296,7 +336,7 @@ pkt_work_mtr(struct rte_mbuf *mbuf,
        drop_mask = MTR_TRTCM_DATA_POLICER_ACTION_DROP_GET(data, color_meter);
        color_policer =
                MTR_TRTCM_DATA_POLICER_ACTION_COLOR_GET(data, color_meter);
-       *sched_ptr = MBUF_SCHED_COLOR(sched, color_policer);
+       rte_mbuf_sched_color_set(mbuf, (uint8_t)color_policer);
 
        return drop_mask;
 }
@@ -318,9 +358,8 @@ tm_cfg_check(struct rte_table_action_tm_config *tm)
 }
 
 struct tm_data {
-       uint16_t queue_tc_color;
-       uint16_t subport;
-       uint32_t pipe;
+       uint32_t queue_id;
+       uint32_t reserved;
 } __attribute__((__packed__));
 
 static int
@@ -347,9 +386,9 @@ tm_apply(struct tm_data *data,
                return status;
 
        /* Apply */
-       data->queue_tc_color = 0;
-       data->subport = (uint16_t) p->subport_id;
-       data->pipe = p->pipe_id;
+       data->queue_id = p->subport_id <<
+                               (__builtin_ctz(cfg->n_pipes_per_subport) + 4) |
+                               p->pipe_id << 4;
 
        return 0;
 }
@@ -361,12 +400,11 @@ pkt_work_tm(struct rte_mbuf *mbuf,
        uint32_t dscp)
 {
        struct dscp_table_entry_data *dscp_entry = &dscp_table->entry[dscp];
-       struct tm_data *sched_ptr = (struct tm_data *) &mbuf->hash.sched;
-       struct tm_data sched;
-
-       sched = *data;
-       sched.queue_tc_color = dscp_entry->queue_tc_color;
-       *sched_ptr = sched;
+       uint32_t queue_id = data->queue_id |
+                               (dscp_entry->tc << 2) |
+                               dscp_entry->tc_queue;
+       rte_mbuf_sched_set(mbuf, queue_id, dscp_entry->tc,
+                               (uint8_t)dscp_entry->color);
 }
 
 /**
@@ -381,6 +419,8 @@ encap_valid(enum rte_table_action_encap_type encap)
        case RTE_TABLE_ACTION_ENCAP_QINQ:
        case RTE_TABLE_ACTION_ENCAP_MPLS:
        case RTE_TABLE_ACTION_ENCAP_PPPOE:
+       case RTE_TABLE_ACTION_ENCAP_VXLAN:
+       case RTE_TABLE_ACTION_ENCAP_QINQ_PPPOE:
                return 1;
        default:
                return 0;
@@ -398,7 +438,7 @@ encap_cfg_check(struct rte_table_action_encap_config *encap)
 }
 
 struct encap_ether_data {
-       struct ether_hdr ether;
+       struct rte_ether_hdr ether;
 } __attribute__((__packed__));
 
 #define VLAN(pcp, dei, vid)                                \
@@ -407,14 +447,14 @@ struct encap_ether_data {
        (((uint64_t)(vid)) & 0xFFFLLU))                    \
 
 struct encap_vlan_data {
-       struct ether_hdr ether;
-       struct vlan_hdr vlan;
+       struct rte_ether_hdr ether;
+       struct rte_vlan_hdr vlan;
 } __attribute__((__packed__));
 
 struct encap_qinq_data {
-       struct ether_hdr ether;
-       struct vlan_hdr svlan;
-       struct vlan_hdr cvlan;
+       struct rte_ether_hdr ether;
+       struct rte_vlan_hdr svlan;
+       struct rte_vlan_hdr cvlan;
 } __attribute__((__packed__));
 
 #define ETHER_TYPE_MPLS_UNICAST                            0x8847
@@ -428,13 +468,11 @@ struct encap_qinq_data {
        (((uint64_t)(ttl)) & 0xFFLLU)))
 
 struct encap_mpls_data {
-       struct ether_hdr ether;
+       struct rte_ether_hdr ether;
        uint32_t mpls[RTE_TABLE_ACTION_MPLS_LABELS_MAX];
        uint32_t mpls_count;
 } __attribute__((__packed__));
 
-#define ETHER_TYPE_PPPOE_SESSION                           0x8864
-
 #define PPP_PROTOCOL_IP                                    0x0021
 
 struct pppoe_ppp_hdr {
@@ -445,7 +483,46 @@ struct pppoe_ppp_hdr {
 } __attribute__((__packed__));
 
 struct encap_pppoe_data {
-       struct ether_hdr ether;
+       struct rte_ether_hdr ether;
+       struct pppoe_ppp_hdr pppoe_ppp;
+} __attribute__((__packed__));
+
+#define IP_PROTO_UDP                                       17
+
+struct encap_vxlan_ipv4_data {
+       struct rte_ether_hdr ether;
+       struct ipv4_hdr ipv4;
+       struct udp_hdr udp;
+       struct rte_vxlan_hdr vxlan;
+} __attribute__((__packed__));
+
+struct encap_vxlan_ipv4_vlan_data {
+       struct rte_ether_hdr ether;
+       struct rte_vlan_hdr vlan;
+       struct ipv4_hdr ipv4;
+       struct udp_hdr udp;
+       struct rte_vxlan_hdr vxlan;
+} __attribute__((__packed__));
+
+struct encap_vxlan_ipv6_data {
+       struct rte_ether_hdr ether;
+       struct ipv6_hdr ipv6;
+       struct udp_hdr udp;
+       struct rte_vxlan_hdr vxlan;
+} __attribute__((__packed__));
+
+struct encap_vxlan_ipv6_vlan_data {
+       struct rte_ether_hdr ether;
+       struct rte_vlan_hdr vlan;
+       struct ipv6_hdr ipv6;
+       struct udp_hdr udp;
+       struct rte_vxlan_hdr vxlan;
+} __attribute__((__packed__));
+
+struct encap_qinq_pppoe_data {
+       struct rte_ether_hdr ether;
+       struct rte_vlan_hdr svlan;
+       struct rte_vlan_hdr cvlan;
        struct pppoe_ppp_hdr pppoe_ppp;
 } __attribute__((__packed__));
 
@@ -468,6 +545,21 @@ encap_data_size(struct rte_table_action_encap_config *encap)
        case 1LLU << RTE_TABLE_ACTION_ENCAP_PPPOE:
                return sizeof(struct encap_pppoe_data);
 
+       case 1LLU << RTE_TABLE_ACTION_ENCAP_VXLAN:
+               if (encap->vxlan.ip_version)
+                       if (encap->vxlan.vlan)
+                               return sizeof(struct encap_vxlan_ipv4_vlan_data);
+                       else
+                               return sizeof(struct encap_vxlan_ipv4_data);
+               else
+                       if (encap->vxlan.vlan)
+                               return sizeof(struct encap_vxlan_ipv6_vlan_data);
+                       else
+                               return sizeof(struct encap_vxlan_ipv6_data);
+
+       case 1LLU << RTE_TABLE_ACTION_ENCAP_QINQ_PPPOE:
+                       return sizeof(struct encap_qinq_pppoe_data);
+
        default:
                return 0;
        }
@@ -501,6 +593,12 @@ encap_apply_check(struct rte_table_action_encap_params *p,
        case RTE_TABLE_ACTION_ENCAP_PPPOE:
                return 0;
 
+       case RTE_TABLE_ACTION_ENCAP_VXLAN:
+               return 0;
+
+       case RTE_TABLE_ACTION_ENCAP_QINQ_PPPOE:
+               return 0;
+
        default:
                return -EINVAL;
        }
@@ -578,6 +676,38 @@ encap_qinq_apply(void *data,
        return 0;
 }
 
+static int
+encap_qinq_pppoe_apply(void *data,
+       struct rte_table_action_encap_params *p)
+{
+       struct encap_qinq_pppoe_data *d = data;
+
+       /* Ethernet */
+       ether_addr_copy(&p->qinq.ether.da, &d->ether.d_addr);
+       ether_addr_copy(&p->qinq.ether.sa, &d->ether.s_addr);
+       d->ether.ether_type = rte_htons(ETHER_TYPE_VLAN);
+
+       /* SVLAN */
+       d->svlan.vlan_tci = rte_htons(VLAN(p->qinq.svlan.pcp,
+               p->qinq.svlan.dei,
+               p->qinq.svlan.vid));
+       d->svlan.eth_proto = rte_htons(ETHER_TYPE_VLAN);
+
+       /* CVLAN */
+       d->cvlan.vlan_tci = rte_htons(VLAN(p->qinq.cvlan.pcp,
+               p->qinq.cvlan.dei,
+               p->qinq.cvlan.vid));
+       d->cvlan.eth_proto = rte_htons(ETHER_TYPE_PPPOE_SESSION);
+
+       /* PPPoE and PPP*/
+       d->pppoe_ppp.ver_type_code = rte_htons(0x1100);
+       d->pppoe_ppp.session_id = rte_htons(p->qinq_pppoe.pppoe.session_id);
+       d->pppoe_ppp.length = 0; /* not pre-computed */
+       d->pppoe_ppp.protocol = rte_htons(PPP_PROTOCOL_IP);
+
+       return 0;
+}
+
 static int
 encap_mpls_apply(void *data,
        struct rte_table_action_encap_params *p)
@@ -629,6 +759,168 @@ encap_pppoe_apply(void *data,
        return 0;
 }
 
+static int
+encap_vxlan_apply(void *data,
+       struct rte_table_action_encap_params *p,
+       struct rte_table_action_encap_config *cfg)
+{
+       if ((p->vxlan.vxlan.vni > 0xFFFFFF) ||
+               (cfg->vxlan.ip_version && (p->vxlan.ipv4.dscp > 0x3F)) ||
+               (!cfg->vxlan.ip_version && (p->vxlan.ipv6.flow_label > 0xFFFFF)) ||
+               (!cfg->vxlan.ip_version && (p->vxlan.ipv6.dscp > 0x3F)) ||
+               (cfg->vxlan.vlan && (p->vxlan.vlan.vid > 0xFFF)))
+               return -1;
+
+       if (cfg->vxlan.ip_version)
+               if (cfg->vxlan.vlan) {
+                       struct encap_vxlan_ipv4_vlan_data *d = data;
+
+                       /* Ethernet */
+                       ether_addr_copy(&p->vxlan.ether.da, &d->ether.d_addr);
+                       ether_addr_copy(&p->vxlan.ether.sa, &d->ether.s_addr);
+                       d->ether.ether_type = rte_htons(ETHER_TYPE_VLAN);
+
+                       /* VLAN */
+                       d->vlan.vlan_tci = rte_htons(VLAN(p->vxlan.vlan.pcp,
+                               p->vxlan.vlan.dei,
+                               p->vxlan.vlan.vid));
+                       d->vlan.eth_proto = rte_htons(ETHER_TYPE_IPv4);
+
+                       /* IPv4*/
+                       d->ipv4.version_ihl = 0x45;
+                       d->ipv4.type_of_service = p->vxlan.ipv4.dscp << 2;
+                       d->ipv4.total_length = 0; /* not pre-computed */
+                       d->ipv4.packet_id = 0;
+                       d->ipv4.fragment_offset = 0;
+                       d->ipv4.time_to_live = p->vxlan.ipv4.ttl;
+                       d->ipv4.next_proto_id = IP_PROTO_UDP;
+                       d->ipv4.hdr_checksum = 0;
+                       d->ipv4.src_addr = rte_htonl(p->vxlan.ipv4.sa);
+                       d->ipv4.dst_addr = rte_htonl(p->vxlan.ipv4.da);
+
+                       d->ipv4.hdr_checksum = rte_ipv4_cksum(&d->ipv4);
+
+                       /* UDP */
+                       d->udp.src_port = rte_htons(p->vxlan.udp.sp);
+                       d->udp.dst_port = rte_htons(p->vxlan.udp.dp);
+                       d->udp.dgram_len = 0; /* not pre-computed */
+                       d->udp.dgram_cksum = 0;
+
+                       /* VXLAN */
+                       d->vxlan.vx_flags = rte_htonl(0x08000000);
+                       d->vxlan.vx_vni = rte_htonl(p->vxlan.vxlan.vni << 8);
+
+                       return 0;
+               } else {
+                       struct encap_vxlan_ipv4_data *d = data;
+
+                       /* Ethernet */
+                       ether_addr_copy(&p->vxlan.ether.da, &d->ether.d_addr);
+                       ether_addr_copy(&p->vxlan.ether.sa, &d->ether.s_addr);
+                       d->ether.ether_type = rte_htons(ETHER_TYPE_IPv4);
+
+                       /* IPv4*/
+                       d->ipv4.version_ihl = 0x45;
+                       d->ipv4.type_of_service = p->vxlan.ipv4.dscp << 2;
+                       d->ipv4.total_length = 0; /* not pre-computed */
+                       d->ipv4.packet_id = 0;
+                       d->ipv4.fragment_offset = 0;
+                       d->ipv4.time_to_live = p->vxlan.ipv4.ttl;
+                       d->ipv4.next_proto_id = IP_PROTO_UDP;
+                       d->ipv4.hdr_checksum = 0;
+                       d->ipv4.src_addr = rte_htonl(p->vxlan.ipv4.sa);
+                       d->ipv4.dst_addr = rte_htonl(p->vxlan.ipv4.da);
+
+                       d->ipv4.hdr_checksum = rte_ipv4_cksum(&d->ipv4);
+
+                       /* UDP */
+                       d->udp.src_port = rte_htons(p->vxlan.udp.sp);
+                       d->udp.dst_port = rte_htons(p->vxlan.udp.dp);
+                       d->udp.dgram_len = 0; /* not pre-computed */
+                       d->udp.dgram_cksum = 0;
+
+                       /* VXLAN */
+                       d->vxlan.vx_flags = rte_htonl(0x08000000);
+                       d->vxlan.vx_vni = rte_htonl(p->vxlan.vxlan.vni << 8);
+
+                       return 0;
+               }
+       else
+               if (cfg->vxlan.vlan) {
+                       struct encap_vxlan_ipv6_vlan_data *d = data;
+
+                       /* Ethernet */
+                       ether_addr_copy(&p->vxlan.ether.da, &d->ether.d_addr);
+                       ether_addr_copy(&p->vxlan.ether.sa, &d->ether.s_addr);
+                       d->ether.ether_type = rte_htons(ETHER_TYPE_VLAN);
+
+                       /* VLAN */
+                       d->vlan.vlan_tci = rte_htons(VLAN(p->vxlan.vlan.pcp,
+                               p->vxlan.vlan.dei,
+                               p->vxlan.vlan.vid));
+                       d->vlan.eth_proto = rte_htons(ETHER_TYPE_IPv6);
+
+                       /* IPv6*/
+                       d->ipv6.vtc_flow = rte_htonl((6 << 28) |
+                               (p->vxlan.ipv6.dscp << 22) |
+                               p->vxlan.ipv6.flow_label);
+                       d->ipv6.payload_len = 0; /* not pre-computed */
+                       d->ipv6.proto = IP_PROTO_UDP;
+                       d->ipv6.hop_limits = p->vxlan.ipv6.hop_limit;
+                       memcpy(d->ipv6.src_addr,
+                               p->vxlan.ipv6.sa,
+                               sizeof(p->vxlan.ipv6.sa));
+                       memcpy(d->ipv6.dst_addr,
+                               p->vxlan.ipv6.da,
+                               sizeof(p->vxlan.ipv6.da));
+
+                       /* UDP */
+                       d->udp.src_port = rte_htons(p->vxlan.udp.sp);
+                       d->udp.dst_port = rte_htons(p->vxlan.udp.dp);
+                       d->udp.dgram_len = 0; /* not pre-computed */
+                       d->udp.dgram_cksum = 0;
+
+                       /* VXLAN */
+                       d->vxlan.vx_flags = rte_htonl(0x08000000);
+                       d->vxlan.vx_vni = rte_htonl(p->vxlan.vxlan.vni << 8);
+
+                       return 0;
+               } else {
+                       struct encap_vxlan_ipv6_data *d = data;
+
+                       /* Ethernet */
+                       ether_addr_copy(&p->vxlan.ether.da, &d->ether.d_addr);
+                       ether_addr_copy(&p->vxlan.ether.sa, &d->ether.s_addr);
+                       d->ether.ether_type = rte_htons(ETHER_TYPE_IPv6);
+
+                       /* IPv6*/
+                       d->ipv6.vtc_flow = rte_htonl((6 << 28) |
+                               (p->vxlan.ipv6.dscp << 22) |
+                               p->vxlan.ipv6.flow_label);
+                       d->ipv6.payload_len = 0; /* not pre-computed */
+                       d->ipv6.proto = IP_PROTO_UDP;
+                       d->ipv6.hop_limits = p->vxlan.ipv6.hop_limit;
+                       memcpy(d->ipv6.src_addr,
+                               p->vxlan.ipv6.sa,
+                               sizeof(p->vxlan.ipv6.sa));
+                       memcpy(d->ipv6.dst_addr,
+                               p->vxlan.ipv6.da,
+                               sizeof(p->vxlan.ipv6.da));
+
+                       /* UDP */
+                       d->udp.src_port = rte_htons(p->vxlan.udp.sp);
+                       d->udp.dst_port = rte_htons(p->vxlan.udp.dp);
+                       d->udp.dgram_len = 0; /* not pre-computed */
+                       d->udp.dgram_cksum = 0;
+
+                       /* VXLAN */
+                       d->vxlan.vx_flags = rte_htonl(0x08000000);
+                       d->vxlan.vx_vni = rte_htonl(p->vxlan.vxlan.vni << 8);
+
+                       return 0;
+               }
+}
+
 static int
 encap_apply(void *data,
        struct rte_table_action_encap_params *p,
@@ -658,11 +950,34 @@ encap_apply(void *data,
        case RTE_TABLE_ACTION_ENCAP_PPPOE:
                return encap_pppoe_apply(data, p);
 
+       case RTE_TABLE_ACTION_ENCAP_VXLAN:
+               return encap_vxlan_apply(data, p, cfg);
+
+       case RTE_TABLE_ACTION_ENCAP_QINQ_PPPOE:
+               return encap_qinq_pppoe_apply(data, p);
+
        default:
                return -EINVAL;
        }
 }
 
+static __rte_always_inline uint16_t
+encap_vxlan_ipv4_checksum_update(uint16_t cksum0,
+       uint16_t total_length)
+{
+       int32_t cksum1;
+
+       cksum1 = cksum0;
+       cksum1 = ~cksum1 & 0xFFFF;
+
+       /* Add total length (one's complement logic) */
+       cksum1 += total_length;
+       cksum1 = (cksum1 & 0xFFFF) + (cksum1 >> 16);
+       cksum1 = (cksum1 & 0xFFFF) + (cksum1 >> 16);
+
+       return (uint16_t)(~cksum1);
+}
+
 static __rte_always_inline void *
 encap(void *dst, const void *src, size_t n)
 {
@@ -670,6 +985,118 @@ encap(void *dst, const void *src, size_t n)
        return rte_memcpy(dst, src, n);
 }
 
+static __rte_always_inline void
+pkt_work_encap_vxlan_ipv4(struct rte_mbuf *mbuf,
+       struct encap_vxlan_ipv4_data *vxlan_tbl,
+       struct rte_table_action_encap_config *cfg)
+{
+       uint32_t ether_offset = cfg->vxlan.data_offset;
+       void *ether = RTE_MBUF_METADATA_UINT32_PTR(mbuf, ether_offset);
+       struct encap_vxlan_ipv4_data *vxlan_pkt;
+       uint16_t ether_length, ipv4_total_length, ipv4_hdr_cksum, udp_length;
+
+       ether_length = (uint16_t)mbuf->pkt_len;
+       ipv4_total_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr) +
+               sizeof(struct ipv4_hdr));
+       ipv4_hdr_cksum = encap_vxlan_ipv4_checksum_update(vxlan_tbl->ipv4.hdr_checksum,
+               rte_htons(ipv4_total_length));
+       udp_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+
+       vxlan_pkt = encap(ether, vxlan_tbl, sizeof(*vxlan_tbl));
+       vxlan_pkt->ipv4.total_length = rte_htons(ipv4_total_length);
+       vxlan_pkt->ipv4.hdr_checksum = ipv4_hdr_cksum;
+       vxlan_pkt->udp.dgram_len = rte_htons(udp_length);
+
+       mbuf->data_off = ether_offset - (sizeof(struct rte_mbuf) + sizeof(*vxlan_pkt));
+       mbuf->pkt_len = mbuf->data_len = ether_length + sizeof(*vxlan_pkt);
+}
+
+static __rte_always_inline void
+pkt_work_encap_vxlan_ipv4_vlan(struct rte_mbuf *mbuf,
+       struct encap_vxlan_ipv4_vlan_data *vxlan_tbl,
+       struct rte_table_action_encap_config *cfg)
+{
+       uint32_t ether_offset = cfg->vxlan.data_offset;
+       void *ether = RTE_MBUF_METADATA_UINT32_PTR(mbuf, ether_offset);
+       struct encap_vxlan_ipv4_vlan_data *vxlan_pkt;
+       uint16_t ether_length, ipv4_total_length, ipv4_hdr_cksum, udp_length;
+
+       ether_length = (uint16_t)mbuf->pkt_len;
+       ipv4_total_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr) +
+               sizeof(struct ipv4_hdr));
+       ipv4_hdr_cksum = encap_vxlan_ipv4_checksum_update(vxlan_tbl->ipv4.hdr_checksum,
+               rte_htons(ipv4_total_length));
+       udp_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+
+       vxlan_pkt = encap(ether, vxlan_tbl, sizeof(*vxlan_tbl));
+       vxlan_pkt->ipv4.total_length = rte_htons(ipv4_total_length);
+       vxlan_pkt->ipv4.hdr_checksum = ipv4_hdr_cksum;
+       vxlan_pkt->udp.dgram_len = rte_htons(udp_length);
+
+       mbuf->data_off = ether_offset - (sizeof(struct rte_mbuf) + sizeof(*vxlan_pkt));
+       mbuf->pkt_len = mbuf->data_len = ether_length + sizeof(*vxlan_pkt);
+}
+
+static __rte_always_inline void
+pkt_work_encap_vxlan_ipv6(struct rte_mbuf *mbuf,
+       struct encap_vxlan_ipv6_data *vxlan_tbl,
+       struct rte_table_action_encap_config *cfg)
+{
+       uint32_t ether_offset = cfg->vxlan.data_offset;
+       void *ether = RTE_MBUF_METADATA_UINT32_PTR(mbuf, ether_offset);
+       struct encap_vxlan_ipv6_data *vxlan_pkt;
+       uint16_t ether_length, ipv6_payload_length, udp_length;
+
+       ether_length = (uint16_t)mbuf->pkt_len;
+       ipv6_payload_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+       udp_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+
+       vxlan_pkt = encap(ether, vxlan_tbl, sizeof(*vxlan_tbl));
+       vxlan_pkt->ipv6.payload_len = rte_htons(ipv6_payload_length);
+       vxlan_pkt->udp.dgram_len = rte_htons(udp_length);
+
+       mbuf->data_off = ether_offset - (sizeof(struct rte_mbuf) + sizeof(*vxlan_pkt));
+       mbuf->pkt_len = mbuf->data_len = ether_length + sizeof(*vxlan_pkt);
+}
+
+static __rte_always_inline void
+pkt_work_encap_vxlan_ipv6_vlan(struct rte_mbuf *mbuf,
+       struct encap_vxlan_ipv6_vlan_data *vxlan_tbl,
+       struct rte_table_action_encap_config *cfg)
+{
+       uint32_t ether_offset = cfg->vxlan.data_offset;
+       void *ether = RTE_MBUF_METADATA_UINT32_PTR(mbuf, ether_offset);
+       struct encap_vxlan_ipv6_vlan_data *vxlan_pkt;
+       uint16_t ether_length, ipv6_payload_length, udp_length;
+
+       ether_length = (uint16_t)mbuf->pkt_len;
+       ipv6_payload_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+       udp_length = ether_length +
+               (sizeof(struct rte_vxlan_hdr) +
+               sizeof(struct udp_hdr));
+
+       vxlan_pkt = encap(ether, vxlan_tbl, sizeof(*vxlan_tbl));
+       vxlan_pkt->ipv6.payload_len = rte_htons(ipv6_payload_length);
+       vxlan_pkt->udp.dgram_len = rte_htons(udp_length);
+
+       mbuf->data_off = ether_offset - (sizeof(struct rte_mbuf) + sizeof(*vxlan_pkt));
+       mbuf->pkt_len = mbuf->data_len = ether_length + sizeof(*vxlan_pkt);
+}
+
 static __rte_always_inline void
 pkt_work_encap(struct rte_mbuf *mbuf,
        void *data,
@@ -706,7 +1133,7 @@ pkt_work_encap(struct rte_mbuf *mbuf,
        case 1LLU << RTE_TABLE_ACTION_ENCAP_MPLS:
        {
                struct encap_mpls_data *mpls = data;
-               size_t size = sizeof(struct ether_hdr) +
+               size_t size = sizeof(struct rte_ether_hdr) +
                        mpls->mpls_count * 4;
 
                encap(ip, data, size);
@@ -727,6 +1154,32 @@ pkt_work_encap(struct rte_mbuf *mbuf,
                break;
        }
 
+       case 1LLU << RTE_TABLE_ACTION_ENCAP_QINQ_PPPOE:
+       {
+               struct encap_qinq_pppoe_data *qinq_pppoe =
+                       encap(ip, data, sizeof(struct encap_qinq_pppoe_data));
+               qinq_pppoe->pppoe_ppp.length = rte_htons(total_length + 2);
+               mbuf->data_off = ip_offset - (sizeof(struct rte_mbuf) +
+                       sizeof(struct encap_qinq_pppoe_data));
+               mbuf->pkt_len = mbuf->data_len = total_length +
+                       sizeof(struct encap_qinq_pppoe_data);
+               break;
+       }
+
+       case 1LLU << RTE_TABLE_ACTION_ENCAP_VXLAN:
+       {
+               if (cfg->vxlan.ip_version)
+                       if (cfg->vxlan.vlan)
+                               pkt_work_encap_vxlan_ipv4_vlan(mbuf, data, cfg);
+                       else
+                               pkt_work_encap_vxlan_ipv4(mbuf, data, cfg);
+               else
+                       if (cfg->vxlan.vlan)
+                               pkt_work_encap_vxlan_ipv6_vlan(mbuf, data, cfg);
+                       else
+                               pkt_work_encap_vxlan_ipv6(mbuf, data, cfg);
+       }
+
        default:
                break;
        }
@@ -1148,6 +1601,583 @@ pkt_work_stats(struct stats_data *data,
        data->n_bytes += total_length;
 }
 
+/**
+ * RTE_TABLE_ACTION_TIME
+ */
+struct time_data {
+       uint64_t time;
+} __attribute__((__packed__));
+
+static int
+time_apply(struct time_data *data,
+       struct rte_table_action_time_params *p)
+{
+       data->time = p->time;
+       return 0;
+}
+
+static __rte_always_inline void
+pkt_work_time(struct time_data *data,
+       uint64_t time)
+{
+       data->time = time;
+}
+
+
+/**
+ * RTE_TABLE_ACTION_CRYPTO
+ */
+
+#define CRYPTO_OP_MASK_CIPHER  0x1
+#define CRYPTO_OP_MASK_AUTH    0x2
+#define CRYPTO_OP_MASK_AEAD    0x4
+
+struct crypto_op_sym_iv_aad {
+       struct rte_crypto_op op;
+       struct rte_crypto_sym_op sym_op;
+       union {
+               struct {
+                       uint8_t cipher_iv[
+                               RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX];
+                       uint8_t auth_iv[
+                               RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX];
+               } cipher_auth;
+
+               struct {
+                       uint8_t iv[RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX];
+                       uint8_t aad[RTE_TABLE_ACTION_SYM_CRYPTO_AAD_SIZE_MAX];
+               } aead_iv_aad;
+
+       } iv_aad;
+};
+
+struct sym_crypto_data {
+
+       union {
+               struct {
+
+                       /** Length of cipher iv. */
+                       uint16_t cipher_iv_len;
+
+                       /** Offset from start of IP header to the cipher iv. */
+                       uint16_t cipher_iv_data_offset;
+
+                       /** Length of cipher iv to be updated in the mbuf. */
+                       uint16_t cipher_iv_update_len;
+
+                       /** Offset from start of IP header to the auth iv. */
+                       uint16_t auth_iv_data_offset;
+
+                       /** Length of auth iv in the mbuf. */
+                       uint16_t auth_iv_len;
+
+                       /** Length of auth iv to be updated in the mbuf. */
+                       uint16_t auth_iv_update_len;
+
+               } cipher_auth;
+               struct {
+
+                       /** Length of iv. */
+                       uint16_t iv_len;
+
+                       /** Offset from start of IP header to the aead iv. */
+                       uint16_t iv_data_offset;
+
+                       /** Length of iv to be updated in the mbuf. */
+                       uint16_t iv_update_len;
+
+                       /** Length of aad */
+                       uint16_t aad_len;
+
+                       /** Offset from start of IP header to the aad. */
+                       uint16_t aad_data_offset;
+
+                       /** Length of aad to updated in the mbuf. */
+                       uint16_t aad_update_len;
+
+               } aead;
+       };
+
+       /** Offset from start of IP header to the data. */
+       uint16_t data_offset;
+
+       /** Digest length. */
+       uint16_t digest_len;
+
+       /** block size */
+       uint16_t block_size;
+
+       /** Mask of crypto operation */
+       uint16_t op_mask;
+
+       /** Session pointer. */
+       struct rte_cryptodev_sym_session *session;
+
+       /** Direction of crypto, encrypt or decrypt */
+       uint16_t direction;
+
+       /** Private data size to store cipher iv / aad. */
+       uint8_t iv_aad_data[32];
+
+} __attribute__((__packed__));
+
+static int
+sym_crypto_cfg_check(struct rte_table_action_sym_crypto_config *cfg)
+{
+       if (!rte_cryptodev_pmd_is_valid_dev(cfg->cryptodev_id))
+               return -EINVAL;
+       if (cfg->mp_create == NULL || cfg->mp_init == NULL)
+               return -EINVAL;
+
+       return 0;
+}
+
+static int
+get_block_size(const struct rte_crypto_sym_xform *xform, uint8_t cdev_id)
+{
+       struct rte_cryptodev_info dev_info;
+       const struct rte_cryptodev_capabilities *cap;
+       uint32_t i;
+
+       rte_cryptodev_info_get(cdev_id, &dev_info);
+
+       for (i = 0; dev_info.capabilities[i].op != RTE_CRYPTO_OP_TYPE_UNDEFINED;
+                       i++) {
+               cap = &dev_info.capabilities[i];
+
+               if (cap->sym.xform_type != xform->type)
+                       continue;
+
+               if ((xform->type == RTE_CRYPTO_SYM_XFORM_CIPHER) &&
+                               (cap->sym.cipher.algo == xform->cipher.algo))
+                       return cap->sym.cipher.block_size;
+
+               if ((xform->type == RTE_CRYPTO_SYM_XFORM_AEAD) &&
+                               (cap->sym.aead.algo == xform->aead.algo))
+                       return cap->sym.aead.block_size;
+
+               if (xform->type == RTE_CRYPTO_SYM_XFORM_NOT_SPECIFIED)
+                       break;
+       }
+
+       return -1;
+}
+
+static int
+sym_crypto_apply(struct sym_crypto_data *data,
+       struct rte_table_action_sym_crypto_config *cfg,
+       struct rte_table_action_sym_crypto_params *p)
+{
+       const struct rte_crypto_cipher_xform *cipher_xform = NULL;
+       const struct rte_crypto_auth_xform *auth_xform = NULL;
+       const struct rte_crypto_aead_xform *aead_xform = NULL;
+       struct rte_crypto_sym_xform *xform = p->xform;
+       struct rte_cryptodev_sym_session *session;
+       int ret;
+
+       memset(data, 0, sizeof(*data));
+
+       while (xform) {
+               if (xform->type == RTE_CRYPTO_SYM_XFORM_CIPHER) {
+                       cipher_xform = &xform->cipher;
+
+                       if (cipher_xform->iv.length >
+                               RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX)
+                               return -ENOMEM;
+                       if (cipher_xform->iv.offset !=
+                                       RTE_TABLE_ACTION_SYM_CRYPTO_IV_OFFSET)
+                               return -EINVAL;
+
+                       ret = get_block_size(xform, cfg->cryptodev_id);
+                       if (ret < 0)
+                               return -1;
+                       data->block_size = (uint16_t)ret;
+                       data->op_mask |= CRYPTO_OP_MASK_CIPHER;
+
+                       data->cipher_auth.cipher_iv_len =
+                                       cipher_xform->iv.length;
+                       data->cipher_auth.cipher_iv_data_offset = (uint16_t)
+                                       p->cipher_auth.cipher_iv_update.offset;
+                       data->cipher_auth.cipher_iv_update_len = (uint16_t)
+                                       p->cipher_auth.cipher_iv_update.length;
+
+                       rte_memcpy(data->iv_aad_data,
+                                       p->cipher_auth.cipher_iv.val,
+                                       p->cipher_auth.cipher_iv.length);
+
+                       data->direction = cipher_xform->op;
+
+               } else if (xform->type == RTE_CRYPTO_SYM_XFORM_AUTH) {
+                       auth_xform = &xform->auth;
+                       if (auth_xform->iv.length >
+                               RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX)
+                               return -ENOMEM;
+                       data->op_mask |= CRYPTO_OP_MASK_AUTH;
+
+                       data->cipher_auth.auth_iv_len = auth_xform->iv.length;
+                       data->cipher_auth.auth_iv_data_offset = (uint16_t)
+                                       p->cipher_auth.auth_iv_update.offset;
+                       data->cipher_auth.auth_iv_update_len = (uint16_t)
+                                       p->cipher_auth.auth_iv_update.length;
+                       data->digest_len = auth_xform->digest_length;
+
+                       data->direction = (auth_xform->op ==
+                                       RTE_CRYPTO_AUTH_OP_GENERATE) ?
+                                       RTE_CRYPTO_CIPHER_OP_ENCRYPT :
+                                       RTE_CRYPTO_CIPHER_OP_DECRYPT;
+
+               } else if (xform->type == RTE_CRYPTO_SYM_XFORM_AEAD) {
+                       aead_xform = &xform->aead;
+
+                       if ((aead_xform->iv.length >
+                               RTE_TABLE_ACTION_SYM_CRYPTO_IV_SIZE_MAX) || (
+                               aead_xform->aad_length >
+                               RTE_TABLE_ACTION_SYM_CRYPTO_AAD_SIZE_MAX))
+                               return -EINVAL;
+                       if (aead_xform->iv.offset !=
+                                       RTE_TABLE_ACTION_SYM_CRYPTO_IV_OFFSET)
+                               return -EINVAL;
+
+                       ret = get_block_size(xform, cfg->cryptodev_id);
+                       if (ret < 0)
+                               return -1;
+                       data->block_size = (uint16_t)ret;
+                       data->op_mask |= CRYPTO_OP_MASK_AEAD;
+
+                       data->digest_len = aead_xform->digest_length;
+                       data->aead.iv_len = aead_xform->iv.length;
+                       data->aead.aad_len = aead_xform->aad_length;
+
+                       data->aead.iv_data_offset = (uint16_t)
+                                       p->aead.iv_update.offset;
+                       data->aead.iv_update_len = (uint16_t)
+                                       p->aead.iv_update.length;
+                       data->aead.aad_data_offset = (uint16_t)
+                                       p->aead.aad_update.offset;
+                       data->aead.aad_update_len = (uint16_t)
+                                       p->aead.aad_update.length;
+
+                       rte_memcpy(data->iv_aad_data,
+                                       p->aead.iv.val,
+                                       p->aead.iv.length);
+
+                       rte_memcpy(data->iv_aad_data + p->aead.iv.length,
+                                       p->aead.aad.val,
+                                       p->aead.aad.length);
+
+                       data->direction = (aead_xform->op ==
+                                       RTE_CRYPTO_AEAD_OP_ENCRYPT) ?
+                                       RTE_CRYPTO_CIPHER_OP_ENCRYPT :
+                                       RTE_CRYPTO_CIPHER_OP_DECRYPT;
+               } else
+                       return -EINVAL;
+
+               xform = xform->next;
+       }
+
+       if (auth_xform && auth_xform->iv.length) {
+               if (cipher_xform) {
+                       if (auth_xform->iv.offset !=
+                                       RTE_TABLE_ACTION_SYM_CRYPTO_IV_OFFSET +
+                                       cipher_xform->iv.length)
+                               return -EINVAL;
+
+                       rte_memcpy(data->iv_aad_data + cipher_xform->iv.length,
+                                       p->cipher_auth.auth_iv.val,
+                                       p->cipher_auth.auth_iv.length);
+               } else {
+                       rte_memcpy(data->iv_aad_data,
+                                       p->cipher_auth.auth_iv.val,
+                                       p->cipher_auth.auth_iv.length);
+               }
+       }
+
+       session = rte_cryptodev_sym_session_create(cfg->mp_create);
+       if (!session)
+               return -ENOMEM;
+
+       ret = rte_cryptodev_sym_session_init(cfg->cryptodev_id, session,
+                       p->xform, cfg->mp_init);
+       if (ret < 0) {
+               rte_cryptodev_sym_session_free(session);
+               return ret;
+       }
+
+       data->data_offset = (uint16_t)p->data_offset;
+       data->session = session;
+
+       return 0;
+}
+
+static __rte_always_inline uint64_t
+pkt_work_sym_crypto(struct rte_mbuf *mbuf, struct sym_crypto_data *data,
+               struct rte_table_action_sym_crypto_config *cfg,
+               uint16_t ip_offset)
+{
+       struct crypto_op_sym_iv_aad *crypto_op = (struct crypto_op_sym_iv_aad *)
+                       RTE_MBUF_METADATA_UINT8_PTR(mbuf, cfg->op_offset);
+       struct rte_crypto_op *op = &crypto_op->op;
+       struct rte_crypto_sym_op *sym = op->sym;
+       uint32_t pkt_offset = sizeof(*mbuf) + mbuf->data_off;
+       uint32_t payload_len = pkt_offset + mbuf->data_len - data->data_offset;
+
+       op->type = RTE_CRYPTO_OP_TYPE_SYMMETRIC;
+       op->sess_type = RTE_CRYPTO_OP_WITH_SESSION;
+       op->phys_addr = mbuf->buf_iova + cfg->op_offset - sizeof(*mbuf);
+       op->status = RTE_CRYPTO_OP_STATUS_NOT_PROCESSED;
+       sym->m_src = mbuf;
+       sym->m_dst = NULL;
+       sym->session = data->session;
+
+       /** pad the packet */
+       if (data->direction == RTE_CRYPTO_CIPHER_OP_ENCRYPT) {
+               uint32_t append_len = RTE_ALIGN_CEIL(payload_len,
+                               data->block_size) - payload_len;
+
+               if (unlikely(rte_pktmbuf_append(mbuf, append_len +
+                               data->digest_len) == NULL))
+                       return 1;
+
+               payload_len += append_len;
+       } else
+               payload_len -= data->digest_len;
+
+       if (data->op_mask & CRYPTO_OP_MASK_CIPHER) {
+               /** prepare cipher op */
+               uint8_t *iv = crypto_op->iv_aad.cipher_auth.cipher_iv;
+
+               sym->cipher.data.length = payload_len;
+               sym->cipher.data.offset = data->data_offset - pkt_offset;
+
+               if (data->cipher_auth.cipher_iv_update_len) {
+                       uint8_t *pkt_iv = RTE_MBUF_METADATA_UINT8_PTR(mbuf,
+                               data->cipher_auth.cipher_iv_data_offset
+                               + ip_offset);
+
+                       /** For encryption, update the pkt iv field, otherwise
+                        *  update the iv_aad_field
+                        **/
+                       if (data->direction == RTE_CRYPTO_CIPHER_OP_ENCRYPT)
+                               rte_memcpy(pkt_iv, data->iv_aad_data,
+                                       data->cipher_auth.cipher_iv_update_len);
+                       else
+                               rte_memcpy(data->iv_aad_data, pkt_iv,
+                                       data->cipher_auth.cipher_iv_update_len);
+               }
+
+               /** write iv */
+               rte_memcpy(iv, data->iv_aad_data,
+                               data->cipher_auth.cipher_iv_len);
+       }
+
+       if (data->op_mask & CRYPTO_OP_MASK_AUTH) {
+               /** authentication always start from IP header. */
+               sym->auth.data.offset = ip_offset - pkt_offset;
+               sym->auth.data.length = mbuf->data_len - sym->auth.data.offset -
+                               data->digest_len;
+               sym->auth.digest.data = rte_pktmbuf_mtod_offset(mbuf,
+                               uint8_t *, rte_pktmbuf_pkt_len(mbuf) -
+                               data->digest_len);
+               sym->auth.digest.phys_addr = rte_pktmbuf_iova_offset(mbuf,
+                               rte_pktmbuf_pkt_len(mbuf) - data->digest_len);
+
+               if (data->cipher_auth.auth_iv_update_len) {
+                       uint8_t *pkt_iv = RTE_MBUF_METADATA_UINT8_PTR(mbuf,
+                                       data->cipher_auth.auth_iv_data_offset
+                                       + ip_offset);
+                       uint8_t *data_iv = data->iv_aad_data +
+                                       data->cipher_auth.cipher_iv_len;
+
+                       if (data->direction == RTE_CRYPTO_CIPHER_OP_ENCRYPT)
+                               rte_memcpy(pkt_iv, data_iv,
+                                       data->cipher_auth.auth_iv_update_len);
+                       else
+                               rte_memcpy(data_iv, pkt_iv,
+                                       data->cipher_auth.auth_iv_update_len);
+               }
+
+               if (data->cipher_auth.auth_iv_len) {
+                       /** prepare cipher op */
+                       uint8_t *iv = crypto_op->iv_aad.cipher_auth.auth_iv;
+
+                       rte_memcpy(iv, data->iv_aad_data +
+                                       data->cipher_auth.cipher_iv_len,
+                                       data->cipher_auth.auth_iv_len);
+               }
+       }
+
+       if (data->op_mask & CRYPTO_OP_MASK_AEAD) {
+               uint8_t *iv = crypto_op->iv_aad.aead_iv_aad.iv;
+               uint8_t *aad = crypto_op->iv_aad.aead_iv_aad.aad;
+
+               sym->aead.aad.data = aad;
+               sym->aead.aad.phys_addr = rte_pktmbuf_iova_offset(mbuf,
+                               aad - rte_pktmbuf_mtod(mbuf, uint8_t *));
+               sym->aead.digest.data = rte_pktmbuf_mtod_offset(mbuf,
+                               uint8_t *, rte_pktmbuf_pkt_len(mbuf) -
+                               data->digest_len);
+               sym->aead.digest.phys_addr = rte_pktmbuf_iova_offset(mbuf,
+                               rte_pktmbuf_pkt_len(mbuf) - data->digest_len);
+               sym->aead.data.offset = data->data_offset - pkt_offset;
+               sym->aead.data.length = payload_len;
+
+               if (data->aead.iv_update_len) {
+                       uint8_t *pkt_iv = RTE_MBUF_METADATA_UINT8_PTR(mbuf,
+                                       data->aead.iv_data_offset + ip_offset);
+                       uint8_t *data_iv = data->iv_aad_data;
+
+                       if (data->direction == RTE_CRYPTO_CIPHER_OP_ENCRYPT)
+                               rte_memcpy(pkt_iv, data_iv,
+                                               data->aead.iv_update_len);
+                       else
+                               rte_memcpy(data_iv, pkt_iv,
+                                       data->aead.iv_update_len);
+               }
+
+               rte_memcpy(iv, data->iv_aad_data, data->aead.iv_len);
+
+               if (data->aead.aad_update_len) {
+                       uint8_t *pkt_aad = RTE_MBUF_METADATA_UINT8_PTR(mbuf,
+                                       data->aead.aad_data_offset + ip_offset);
+                       uint8_t *data_aad = data->iv_aad_data +
+                                       data->aead.iv_len;
+
+                       if (data->direction == RTE_CRYPTO_CIPHER_OP_ENCRYPT)
+                               rte_memcpy(pkt_aad, data_aad,
+                                               data->aead.iv_update_len);
+                       else
+                               rte_memcpy(data_aad, pkt_aad,
+                                       data->aead.iv_update_len);
+               }
+
+               rte_memcpy(aad, data->iv_aad_data + data->aead.iv_len,
+                                       data->aead.aad_len);
+       }
+
+       return 0;
+}
+
+/**
+ * RTE_TABLE_ACTION_TAG
+ */
+struct tag_data {
+       uint32_t tag;
+} __attribute__((__packed__));
+
+static int
+tag_apply(struct tag_data *data,
+       struct rte_table_action_tag_params *p)
+{
+       data->tag = p->tag;
+       return 0;
+}
+
+static __rte_always_inline void
+pkt_work_tag(struct rte_mbuf *mbuf,
+       struct tag_data *data)
+{
+       mbuf->hash.fdir.hi = data->tag;
+       mbuf->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
+}
+
+static __rte_always_inline void
+pkt4_work_tag(struct rte_mbuf *mbuf0,
+       struct rte_mbuf *mbuf1,
+       struct rte_mbuf *mbuf2,
+       struct rte_mbuf *mbuf3,
+       struct tag_data *data0,
+       struct tag_data *data1,
+       struct tag_data *data2,
+       struct tag_data *data3)
+{
+       mbuf0->hash.fdir.hi = data0->tag;
+       mbuf1->hash.fdir.hi = data1->tag;
+       mbuf2->hash.fdir.hi = data2->tag;
+       mbuf3->hash.fdir.hi = data3->tag;
+
+       mbuf0->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
+       mbuf1->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
+       mbuf2->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
+       mbuf3->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
+}
+
+/**
+ * RTE_TABLE_ACTION_DECAP
+ */
+struct decap_data {
+       uint16_t n;
+} __attribute__((__packed__));
+
+static int
+decap_apply(struct decap_data *data,
+       struct rte_table_action_decap_params *p)
+{
+       data->n = p->n;
+       return 0;
+}
+
+static __rte_always_inline void
+pkt_work_decap(struct rte_mbuf *mbuf,
+       struct decap_data *data)
+{
+       uint16_t data_off = mbuf->data_off;
+       uint16_t data_len = mbuf->data_len;
+       uint32_t pkt_len = mbuf->pkt_len;
+       uint16_t n = data->n;
+
+       mbuf->data_off = data_off + n;
+       mbuf->data_len = data_len - n;
+       mbuf->pkt_len = pkt_len - n;
+}
+
+static __rte_always_inline void
+pkt4_work_decap(struct rte_mbuf *mbuf0,
+       struct rte_mbuf *mbuf1,
+       struct rte_mbuf *mbuf2,
+       struct rte_mbuf *mbuf3,
+       struct decap_data *data0,
+       struct decap_data *data1,
+       struct decap_data *data2,
+       struct decap_data *data3)
+{
+       uint16_t data_off0 = mbuf0->data_off;
+       uint16_t data_len0 = mbuf0->data_len;
+       uint32_t pkt_len0 = mbuf0->pkt_len;
+
+       uint16_t data_off1 = mbuf1->data_off;
+       uint16_t data_len1 = mbuf1->data_len;
+       uint32_t pkt_len1 = mbuf1->pkt_len;
+
+       uint16_t data_off2 = mbuf2->data_off;
+       uint16_t data_len2 = mbuf2->data_len;
+       uint32_t pkt_len2 = mbuf2->pkt_len;
+
+       uint16_t data_off3 = mbuf3->data_off;
+       uint16_t data_len3 = mbuf3->data_len;
+       uint32_t pkt_len3 = mbuf3->pkt_len;
+
+       uint16_t n0 = data0->n;
+       uint16_t n1 = data1->n;
+       uint16_t n2 = data2->n;
+       uint16_t n3 = data3->n;
+
+       mbuf0->data_off = data_off0 + n0;
+       mbuf0->data_len = data_len0 - n0;
+       mbuf0->pkt_len = pkt_len0 - n0;
+
+       mbuf1->data_off = data_off1 + n1;
+       mbuf1->data_len = data_len1 - n1;
+       mbuf1->pkt_len = pkt_len1 - n1;
+
+       mbuf2->data_off = data_off2 + n2;
+       mbuf2->data_len = data_len2 - n2;
+       mbuf2->pkt_len = pkt_len2 - n2;
+
+       mbuf3->data_off = data_off3 + n3;
+       mbuf3->data_len = data_len3 - n3;
+       mbuf3->pkt_len = pkt_len3 - n3;
+}
+
 /**
  * Action profile
  */
@@ -1156,12 +2186,17 @@ action_valid(enum rte_table_action_type action)
 {
        switch (action) {
        case RTE_TABLE_ACTION_FWD:
+       case RTE_TABLE_ACTION_LB:
        case RTE_TABLE_ACTION_MTR:
        case RTE_TABLE_ACTION_TM:
        case RTE_TABLE_ACTION_ENCAP:
        case RTE_TABLE_ACTION_NAT:
        case RTE_TABLE_ACTION_TTL:
        case RTE_TABLE_ACTION_STATS:
+       case RTE_TABLE_ACTION_TIME:
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+       case RTE_TABLE_ACTION_TAG:
+       case RTE_TABLE_ACTION_DECAP:
                return 1;
        default:
                return 0;
@@ -1174,18 +2209,22 @@ action_valid(enum rte_table_action_type action)
 struct ap_config {
        uint64_t action_mask;
        struct rte_table_action_common_config common;
+       struct rte_table_action_lb_config lb;
        struct rte_table_action_mtr_config mtr;
        struct rte_table_action_tm_config tm;
        struct rte_table_action_encap_config encap;
        struct rte_table_action_nat_config nat;
        struct rte_table_action_ttl_config ttl;
        struct rte_table_action_stats_config stats;
+       struct rte_table_action_sym_crypto_config sym_crypto;
 };
 
 static size_t
 action_cfg_size(enum rte_table_action_type action)
 {
        switch (action) {
+       case RTE_TABLE_ACTION_LB:
+               return sizeof(struct rte_table_action_lb_config);
        case RTE_TABLE_ACTION_MTR:
                return sizeof(struct rte_table_action_mtr_config);
        case RTE_TABLE_ACTION_TM:
@@ -1198,6 +2237,8 @@ action_cfg_size(enum rte_table_action_type action)
                return sizeof(struct rte_table_action_ttl_config);
        case RTE_TABLE_ACTION_STATS:
                return sizeof(struct rte_table_action_stats_config);
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+               return sizeof(struct rte_table_action_sym_crypto_config);
        default:
                return 0;
        }
@@ -1208,6 +2249,9 @@ action_cfg_get(struct ap_config *ap_config,
        enum rte_table_action_type type)
 {
        switch (type) {
+       case RTE_TABLE_ACTION_LB:
+               return &ap_config->lb;
+
        case RTE_TABLE_ACTION_MTR:
                return &ap_config->mtr;
 
@@ -1226,6 +2270,8 @@ action_cfg_get(struct ap_config *ap_config,
        case RTE_TABLE_ACTION_STATS:
                return &ap_config->stats;
 
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+               return &ap_config->sym_crypto;
        default:
                return NULL;
        }
@@ -1257,6 +2303,9 @@ action_data_size(enum rte_table_action_type action,
        case RTE_TABLE_ACTION_FWD:
                return sizeof(struct fwd_data);
 
+       case RTE_TABLE_ACTION_LB:
+               return sizeof(struct lb_data);
+
        case RTE_TABLE_ACTION_MTR:
                return mtr_data_size(&ap_config->mtr);
 
@@ -1276,6 +2325,18 @@ action_data_size(enum rte_table_action_type action,
        case RTE_TABLE_ACTION_STATS:
                return sizeof(struct stats_data);
 
+       case RTE_TABLE_ACTION_TIME:
+               return sizeof(struct time_data);
+
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+               return (sizeof(struct sym_crypto_data));
+
+       case RTE_TABLE_ACTION_TAG:
+               return sizeof(struct tag_data);
+
+       case RTE_TABLE_ACTION_DECAP:
+               return sizeof(struct decap_data);
+
        default:
                return 0;
        }
@@ -1347,6 +2408,10 @@ rte_table_action_profile_action_register(struct rte_table_action_profile *profil
                return -EINVAL;
 
        switch (type) {
+       case RTE_TABLE_ACTION_LB:
+               status = lb_cfg_check(action_config);
+               break;
+
        case RTE_TABLE_ACTION_MTR:
                status = mtr_cfg_check(action_config);
                break;
@@ -1371,6 +2436,10 @@ rte_table_action_profile_action_register(struct rte_table_action_profile *profil
                status = stats_cfg_check(action_config);
                break;
 
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+               status = sym_crypto_cfg_check(action_config);
+               break;
+
        default:
                status = 0;
                break;
@@ -1481,6 +2550,10 @@ rte_table_action_apply(struct rte_table_action *action,
                return fwd_apply(action_data,
                        action_params);
 
+       case RTE_TABLE_ACTION_LB:
+               return lb_apply(action_data,
+                       action_params);
+
        case RTE_TABLE_ACTION_MTR:
                return mtr_apply(action_data,
                        action_params,
@@ -1512,6 +2585,23 @@ rte_table_action_apply(struct rte_table_action *action,
                return stats_apply(action_data,
                        action_params);
 
+       case RTE_TABLE_ACTION_TIME:
+               return time_apply(action_data,
+                       action_params);
+
+       case RTE_TABLE_ACTION_SYM_CRYPTO:
+               return sym_crypto_apply(action_data,
+                               &action->cfg.sym_crypto,
+                               action_params);
+
+       case RTE_TABLE_ACTION_TAG:
+               return tag_apply(action_data,
+                       action_params);
+
+       case RTE_TABLE_ACTION_DECAP:
+               return decap_apply(action_data,
+                       action_params);
+
        default:
                return -EINVAL;
        }
@@ -1537,17 +2627,13 @@ rte_table_action_dscp_table_update(struct rte_table_action *action,
                        &action->dscp_table.entry[i];
                struct rte_table_action_dscp_table_entry *entry =
                        &table->entry[i];
-               uint16_t queue_tc_color =
-                       MBUF_SCHED_QUEUE_TC_COLOR(entry->tc_queue_id,
-                               entry->tc_id,
-                               entry->color);
 
                if ((dscp_mask & (1LLU << i)) == 0)
                        continue;
 
                data->color = entry->color;
                data->tc = entry->tc_id;
-               data->queue_tc_color = queue_tc_color;
+               data->tc_queue = entry->tc_queue_id;
        }
 
        return 0;
@@ -1645,14 +2731,14 @@ rte_table_action_meter_read(struct rte_table_action *action,
                        if ((tc_mask & (1 << i)) == 0)
                                continue;
 
-                       dst->n_packets[e_RTE_METER_GREEN] =
-                               mtr_trtcm_data_stats_get(src, e_RTE_METER_GREEN);
+                       dst->n_packets[RTE_COLOR_GREEN] =
+                               mtr_trtcm_data_stats_get(src, RTE_COLOR_GREEN);
 
-                       dst->n_packets[e_RTE_METER_YELLOW] =
-                               mtr_trtcm_data_stats_get(src, e_RTE_METER_YELLOW);
+                       dst->n_packets[RTE_COLOR_YELLOW] =
+                               mtr_trtcm_data_stats_get(src, RTE_COLOR_YELLOW);
 
-                       dst->n_packets[e_RTE_METER_RED] =
-                               mtr_trtcm_data_stats_get(src, e_RTE_METER_RED);
+                       dst->n_packets[RTE_COLOR_RED] =
+                               mtr_trtcm_data_stats_get(src, RTE_COLOR_RED);
 
                        dst->n_packets_valid = 1;
                        dst->n_bytes_valid = 0;
@@ -1669,9 +2755,9 @@ rte_table_action_meter_read(struct rte_table_action *action,
                        if ((tc_mask & (1 << i)) == 0)
                                continue;
 
-                       mtr_trtcm_data_stats_reset(src, e_RTE_METER_GREEN);
-                       mtr_trtcm_data_stats_reset(src, e_RTE_METER_YELLOW);
-                       mtr_trtcm_data_stats_reset(src, e_RTE_METER_RED);
+                       mtr_trtcm_data_stats_reset(src, RTE_COLOR_GREEN);
+                       mtr_trtcm_data_stats_reset(src, RTE_COLOR_YELLOW);
+                       mtr_trtcm_data_stats_reset(src, RTE_COLOR_RED);
                }
 
 
@@ -1741,6 +2827,48 @@ rte_table_action_stats_read(struct rte_table_action *action,
        return 0;
 }
 
+int
+rte_table_action_time_read(struct rte_table_action *action,
+       void *data,
+       uint64_t *timestamp)
+{
+       struct time_data *time_data;
+
+       /* Check input arguments */
+       if ((action == NULL) ||
+               ((action->cfg.action_mask &
+               (1LLU << RTE_TABLE_ACTION_TIME)) == 0) ||
+               (data == NULL) ||
+               (timestamp == NULL))
+               return -EINVAL;
+
+       time_data = action_data_get(data, action, RTE_TABLE_ACTION_TIME);
+
+       /* Read */
+       *timestamp = time_data->time;
+
+       return 0;
+}
+
+struct rte_cryptodev_sym_session *
+rte_table_action_crypto_sym_session_get(struct rte_table_action *action,
+       void *data)
+{
+       struct sym_crypto_data *sym_crypto_data;
+
+       /* Check input arguments */
+       if ((action == NULL) ||
+               ((action->cfg.action_mask &
+               (1LLU << RTE_TABLE_ACTION_SYM_CRYPTO)) == 0) ||
+               (data == NULL))
+               return NULL;
+
+       sym_crypto_data = action_data_get(data, action,
+                       RTE_TABLE_ACTION_SYM_CRYPTO);
+
+       return sym_crypto_data->session;
+}
+
 static __rte_always_inline uint64_t
 pkt_work(struct rte_mbuf *mbuf,
        struct rte_pipeline_table_entry *table_entry,
@@ -1769,6 +2897,14 @@ pkt_work(struct rte_mbuf *mbuf,
                        rte_ntohs(hdr->payload_len) + sizeof(struct ipv6_hdr);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_LB)) {
+               void *data =
+                       action_data_get(table_entry, action, RTE_TABLE_ACTION_LB);
+
+               pkt_work_lb(mbuf,
+                       data,
+                       &cfg->lb);
+       }
        if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_MTR)) {
                void *data =
                        action_data_get(table_entry, action, RTE_TABLE_ACTION_MTR);
@@ -1792,6 +2928,14 @@ pkt_work(struct rte_mbuf *mbuf,
                        dscp);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_DECAP)) {
+               void *data = action_data_get(table_entry,
+                       action,
+                       RTE_TABLE_ACTION_DECAP);
+
+               pkt_work_decap(mbuf, data);
+       }
+
        if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_ENCAP)) {
                void *data =
                        action_data_get(table_entry, action, RTE_TABLE_ACTION_ENCAP);
@@ -1831,6 +2975,29 @@ pkt_work(struct rte_mbuf *mbuf,
                pkt_work_stats(data, total_length);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_TIME)) {
+               void *data =
+                       action_data_get(table_entry, action, RTE_TABLE_ACTION_TIME);
+
+               pkt_work_time(data, time);
+       }
+
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_SYM_CRYPTO)) {
+               void *data = action_data_get(table_entry, action,
+                               RTE_TABLE_ACTION_SYM_CRYPTO);
+
+               drop_mask |= pkt_work_sym_crypto(mbuf, data, &cfg->sym_crypto,
+                               ip_offset);
+       }
+
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_TAG)) {
+               void *data = action_data_get(table_entry,
+                       action,
+                       RTE_TABLE_ACTION_TAG);
+
+               pkt_work_tag(mbuf, data);
+       }
+
        return drop_mask;
 }
 
@@ -1901,6 +3068,33 @@ pkt4_work(struct rte_mbuf **mbufs,
                        rte_ntohs(hdr3->payload_len) + sizeof(struct ipv6_hdr);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_LB)) {
+               void *data0 =
+                       action_data_get(table_entry0, action, RTE_TABLE_ACTION_LB);
+               void *data1 =
+                       action_data_get(table_entry1, action, RTE_TABLE_ACTION_LB);
+               void *data2 =
+                       action_data_get(table_entry2, action, RTE_TABLE_ACTION_LB);
+               void *data3 =
+                       action_data_get(table_entry3, action, RTE_TABLE_ACTION_LB);
+
+               pkt_work_lb(mbuf0,
+                       data0,
+                       &cfg->lb);
+
+               pkt_work_lb(mbuf1,
+                       data1,
+                       &cfg->lb);
+
+               pkt_work_lb(mbuf2,
+                       data2,
+                       &cfg->lb);
+
+               pkt_work_lb(mbuf3,
+                       data3,
+                       &cfg->lb);
+       }
+
        if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_MTR)) {
                void *data0 =
                        action_data_get(table_entry0, action, RTE_TABLE_ACTION_MTR);
@@ -1975,6 +3169,24 @@ pkt4_work(struct rte_mbuf **mbufs,
                        dscp3);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_DECAP)) {
+               void *data0 = action_data_get(table_entry0,
+                       action,
+                       RTE_TABLE_ACTION_DECAP);
+               void *data1 = action_data_get(table_entry1,
+                       action,
+                       RTE_TABLE_ACTION_DECAP);
+               void *data2 = action_data_get(table_entry2,
+                       action,
+                       RTE_TABLE_ACTION_DECAP);
+               void *data3 = action_data_get(table_entry3,
+                       action,
+                       RTE_TABLE_ACTION_DECAP);
+
+               pkt4_work_decap(mbuf0, mbuf1, mbuf2, mbuf3,
+                       data0, data1, data2, data3);
+       }
+
        if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_ENCAP)) {
                void *data0 =
                        action_data_get(table_entry0, action, RTE_TABLE_ACTION_ENCAP);
@@ -2076,6 +3288,60 @@ pkt4_work(struct rte_mbuf **mbufs,
                pkt_work_stats(data3, total_length3);
        }
 
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_TIME)) {
+               void *data0 =
+                       action_data_get(table_entry0, action, RTE_TABLE_ACTION_TIME);
+               void *data1 =
+                       action_data_get(table_entry1, action, RTE_TABLE_ACTION_TIME);
+               void *data2 =
+                       action_data_get(table_entry2, action, RTE_TABLE_ACTION_TIME);
+               void *data3 =
+                       action_data_get(table_entry3, action, RTE_TABLE_ACTION_TIME);
+
+               pkt_work_time(data0, time);
+               pkt_work_time(data1, time);
+               pkt_work_time(data2, time);
+               pkt_work_time(data3, time);
+       }
+
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_SYM_CRYPTO)) {
+               void *data0 = action_data_get(table_entry0, action,
+                               RTE_TABLE_ACTION_SYM_CRYPTO);
+               void *data1 = action_data_get(table_entry1, action,
+                               RTE_TABLE_ACTION_SYM_CRYPTO);
+               void *data2 = action_data_get(table_entry2, action,
+                               RTE_TABLE_ACTION_SYM_CRYPTO);
+               void *data3 = action_data_get(table_entry3, action,
+                               RTE_TABLE_ACTION_SYM_CRYPTO);
+
+               drop_mask0 |= pkt_work_sym_crypto(mbuf0, data0, &cfg->sym_crypto,
+                               ip_offset);
+               drop_mask1 |= pkt_work_sym_crypto(mbuf1, data1, &cfg->sym_crypto,
+                               ip_offset);
+               drop_mask2 |= pkt_work_sym_crypto(mbuf2, data2, &cfg->sym_crypto,
+                               ip_offset);
+               drop_mask3 |= pkt_work_sym_crypto(mbuf3, data3, &cfg->sym_crypto,
+                               ip_offset);
+       }
+
+       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_TAG)) {
+               void *data0 = action_data_get(table_entry0,
+                       action,
+                       RTE_TABLE_ACTION_TAG);
+               void *data1 = action_data_get(table_entry1,
+                       action,
+                       RTE_TABLE_ACTION_TAG);
+               void *data2 = action_data_get(table_entry2,
+                       action,
+                       RTE_TABLE_ACTION_TAG);
+               void *data3 = action_data_get(table_entry3,
+                       action,
+                       RTE_TABLE_ACTION_TAG);
+
+               pkt4_work_tag(mbuf0, mbuf1, mbuf2, mbuf3,
+                       data0, data1, data2, data3);
+       }
+
        return drop_mask0 |
                (drop_mask1 << 1) |
                (drop_mask2 << 2) |
@@ -2093,7 +3359,8 @@ ah(struct rte_pipeline *p,
        uint64_t pkts_drop_mask = 0;
        uint64_t time = 0;
 
-       if (cfg->action_mask & (1LLU << RTE_TABLE_ACTION_MTR))
+       if (cfg->action_mask & ((1LLU << RTE_TABLE_ACTION_MTR) |
+               (1LLU << RTE_TABLE_ACTION_TIME)))
                time = rte_rdtsc();
 
        if ((pkts_mask & (pkts_mask + 1)) == 0) {