net/enic: remove deprecated flow director code
authorJohn Daley <johndale@cisco.com>
Thu, 17 Dec 2020 01:37:15 +0000 (17:37 -0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Wed, 13 Jan 2021 17:51:57 +0000 (18:51 +0100)
The Flow Director (FDIR) API was removed in release 20.11.
This patch removes the remainder of the FDIR code in the
PMD.

Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Hyong Youb Kim <hyonkim@cisco.com>
drivers/net/enic/enic.h
drivers/net/enic/enic_clsf.c [deleted file]
drivers/net/enic/enic_flow.c
drivers/net/enic/enic_fm_flow.c
drivers/net/enic/enic_main.c
drivers/net/enic/enic_rxtx_vec_avx2.c
drivers/net/enic/meson.build

index 079f194..4ee7525 100644 (file)
@@ -51,8 +51,6 @@
 /* Special Filter id for non-specific packet flagging. Don't change value */
 #define ENIC_MAGIC_FILTER_ID 0xffff
 
-#define ENICPMD_FDIR_MAX           64
-
 /*
  * Interrupt 0: LSC and errors
  * Interrupt 1: rx queue 0
 #define ENICPMD_LSC_INTR_OFFSET 0
 #define ENICPMD_RXQ_INTR_OFFSET 1
 
-struct enic_fdir_node {
-       struct rte_eth_fdir_filter filter;
-       uint16_t fltr_id;
-       uint16_t rq_index;
-};
-
-struct enic_fdir {
-       struct rte_eth_fdir_stats stats;
-       struct rte_hash *hash;
-       struct enic_fdir_node *nodes[ENICPMD_FDIR_MAX];
-       uint32_t modes;
-       uint32_t types_mask;
-       void (*copy_fltr_fn)(struct filter_v2 *filt,
-                            const struct rte_eth_fdir_input *input,
-                            const struct rte_eth_fdir_masks *masks);
-};
-
 struct enic_soft_stats {
        rte_atomic64_t rx_nombuf;
        rte_atomic64_t rx_packet_errors;
@@ -120,7 +101,6 @@ struct enic {
        bool overlay_offload;
        struct rte_eth_dev *rte_dev;
        struct rte_eth_dev_data *dev_data;
-       struct enic_fdir fdir;
        char bdf_name[ENICPMD_BDF_LENGTH];
        int dev_fd;
        int iommu_group_fd;
@@ -431,8 +411,6 @@ void enic_send_pkt(struct enic *enic, struct vnic_wq *wq,
 
 void enic_post_wq_index(struct vnic_wq *wq);
 int enic_probe(struct enic *enic);
-int enic_clsf_init(struct enic *enic);
-void enic_clsf_destroy(struct enic *enic);
 int enic_fm_init(struct enic *enic);
 void enic_fm_destroy(struct enic *enic);
 void *enic_alloc_consistent(void *priv, size_t size, dma_addr_t *dma_handle,
@@ -457,7 +435,6 @@ int enic_link_update(struct rte_eth_dev *eth_dev);
 bool enic_use_vector_rx_handler(struct rte_eth_dev *eth_dev);
 void enic_pick_rx_handler(struct rte_eth_dev *eth_dev);
 void enic_pick_tx_handler(struct rte_eth_dev *eth_dev);
-void enic_fdir_info(struct enic *enic);
 int enic_vf_representor_init(struct rte_eth_dev *eth_dev, void *init_params);
 int enic_vf_representor_uninit(struct rte_eth_dev *ethdev);
 int enic_fm_allocate_switch_domain(struct enic *pf);
diff --git a/drivers/net/enic/enic_clsf.c b/drivers/net/enic/enic_clsf.c
deleted file mode 100644 (file)
index 1c837a4..0000000
+++ /dev/null
@@ -1,334 +0,0 @@
-/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright 2008-2017 Cisco Systems, Inc.  All rights reserved.
- * Copyright 2007 Nuova Systems, Inc.  All rights reserved.
- */
-
-#include <rte_ethdev_driver.h>
-#include <rte_malloc.h>
-#include <rte_hash.h>
-#include <rte_byteorder.h>
-#include <rte_ip.h>
-#include <rte_tcp.h>
-#include <rte_udp.h>
-#include <rte_sctp.h>
-
-#include "enic_compat.h"
-#include "enic.h"
-#include "wq_enet_desc.h"
-#include "rq_enet_desc.h"
-#include "cq_enet_desc.h"
-#include "vnic_enet.h"
-#include "vnic_dev.h"
-#include "vnic_wq.h"
-#include "vnic_rq.h"
-#include "vnic_cq.h"
-#include "vnic_intr.h"
-#include "vnic_nic.h"
-
-#ifdef RTE_ARCH_X86
-#include <rte_hash_crc.h>
-#define DEFAULT_HASH_FUNC       rte_hash_crc
-#else
-#include <rte_jhash.h>
-#define DEFAULT_HASH_FUNC       rte_jhash
-#endif
-
-#define ENICPMD_CLSF_HASH_ENTRIES       ENICPMD_FDIR_MAX
-
-static void copy_fltr_v1(struct filter_v2 *fltr,
-               const struct rte_eth_fdir_input *input,
-               const struct rte_eth_fdir_masks *masks);
-static void copy_fltr_v2(struct filter_v2 *fltr,
-               const struct rte_eth_fdir_input *input,
-               const struct rte_eth_fdir_masks *masks);
-
-void enic_fdir_info(struct enic *enic)
-{
-       enic->fdir.modes = (uint32_t)RTE_FDIR_MODE_PERFECT;
-       enic->fdir.types_mask  = 1 << RTE_ETH_FLOW_NONFRAG_IPV4_UDP |
-                                1 << RTE_ETH_FLOW_NONFRAG_IPV4_TCP;
-       if (enic->adv_filters) {
-               enic->fdir.types_mask |= 1 << RTE_ETH_FLOW_NONFRAG_IPV4_OTHER |
-                                        1 << RTE_ETH_FLOW_NONFRAG_IPV4_SCTP |
-                                        1 << RTE_ETH_FLOW_NONFRAG_IPV6_UDP |
-                                        1 << RTE_ETH_FLOW_NONFRAG_IPV6_TCP |
-                                        1 << RTE_ETH_FLOW_NONFRAG_IPV6_SCTP |
-                                        1 << RTE_ETH_FLOW_NONFRAG_IPV6_OTHER;
-               enic->fdir.copy_fltr_fn = copy_fltr_v2;
-       } else {
-               enic->fdir.copy_fltr_fn = copy_fltr_v1;
-       }
-}
-
-static void
-enic_set_layer(struct filter_generic_1 *gp, unsigned int flag,
-              enum filter_generic_1_layer layer, void *mask, void *val,
-              unsigned int len)
-{
-       gp->mask_flags |= flag;
-       gp->val_flags |= gp->mask_flags;
-       memcpy(gp->layer[layer].mask, mask, len);
-       memcpy(gp->layer[layer].val, val, len);
-}
-
-/* Copy Flow Director filter to a VIC ipv4 filter (for Cisco VICs
- * without advanced filter support.
- */
-static void
-copy_fltr_v1(struct filter_v2 *fltr, const struct rte_eth_fdir_input *input,
-            __rte_unused const struct rte_eth_fdir_masks *masks)
-{
-       fltr->type = FILTER_IPV4_5TUPLE;
-       fltr->u.ipv4.src_addr = rte_be_to_cpu_32(
-               input->flow.ip4_flow.src_ip);
-       fltr->u.ipv4.dst_addr = rte_be_to_cpu_32(
-               input->flow.ip4_flow.dst_ip);
-       fltr->u.ipv4.src_port = rte_be_to_cpu_16(
-               input->flow.udp4_flow.src_port);
-       fltr->u.ipv4.dst_port = rte_be_to_cpu_16(
-               input->flow.udp4_flow.dst_port);
-
-       if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_TCP)
-               fltr->u.ipv4.protocol = PROTO_TCP;
-       else
-               fltr->u.ipv4.protocol = PROTO_UDP;
-
-       fltr->u.ipv4.flags = FILTER_FIELDS_IPV4_5TUPLE;
-}
-
-/* Copy Flow Director filter to a VIC generic filter (requires advanced
- * filter support.
- */
-static void
-copy_fltr_v2(struct filter_v2 *fltr, const struct rte_eth_fdir_input *input,
-            const struct rte_eth_fdir_masks *masks)
-{
-       struct filter_generic_1 *gp = &fltr->u.generic_1;
-
-       fltr->type = FILTER_DPDK_1;
-       memset(gp, 0, sizeof(*gp));
-
-       if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_UDP) {
-               struct rte_udp_hdr udp_mask, udp_val;
-               memset(&udp_mask, 0, sizeof(udp_mask));
-               memset(&udp_val, 0, sizeof(udp_val));
-
-               if (input->flow.udp4_flow.src_port) {
-                       udp_mask.src_port = masks->src_port_mask;
-                       udp_val.src_port = input->flow.udp4_flow.src_port;
-               }
-               if (input->flow.udp4_flow.dst_port) {
-                       udp_mask.dst_port = masks->dst_port_mask;
-                       udp_val.dst_port = input->flow.udp4_flow.dst_port;
-               }
-
-               enic_set_layer(gp, FILTER_GENERIC_1_UDP, FILTER_GENERIC_1_L4,
-                              &udp_mask, &udp_val, sizeof(struct rte_udp_hdr));
-       } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_TCP) {
-               struct rte_tcp_hdr tcp_mask, tcp_val;
-               memset(&tcp_mask, 0, sizeof(tcp_mask));
-               memset(&tcp_val, 0, sizeof(tcp_val));
-
-               if (input->flow.tcp4_flow.src_port) {
-                       tcp_mask.src_port = masks->src_port_mask;
-                       tcp_val.src_port = input->flow.tcp4_flow.src_port;
-               }
-               if (input->flow.tcp4_flow.dst_port) {
-                       tcp_mask.dst_port = masks->dst_port_mask;
-                       tcp_val.dst_port = input->flow.tcp4_flow.dst_port;
-               }
-
-               enic_set_layer(gp, FILTER_GENERIC_1_TCP, FILTER_GENERIC_1_L4,
-                              &tcp_mask, &tcp_val, sizeof(struct rte_tcp_hdr));
-       } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_SCTP) {
-               struct rte_sctp_hdr sctp_mask, sctp_val;
-               memset(&sctp_mask, 0, sizeof(sctp_mask));
-               memset(&sctp_val, 0, sizeof(sctp_val));
-
-               if (input->flow.sctp4_flow.src_port) {
-                       sctp_mask.src_port = masks->src_port_mask;
-                       sctp_val.src_port = input->flow.sctp4_flow.src_port;
-               }
-               if (input->flow.sctp4_flow.dst_port) {
-                       sctp_mask.dst_port = masks->dst_port_mask;
-                       sctp_val.dst_port = input->flow.sctp4_flow.dst_port;
-               }
-               if (input->flow.sctp4_flow.verify_tag) {
-                       sctp_mask.tag = 0xffffffff;
-                       sctp_val.tag = input->flow.sctp4_flow.verify_tag;
-               }
-
-               /*
-                * Unlike UDP/TCP (FILTER_GENERIC_1_{UDP,TCP}), the firmware
-                * has no "packet is SCTP" flag. Use flag=0 (generic L4) and
-                * manually set proto_id=sctp below.
-                */
-               enic_set_layer(gp, 0, FILTER_GENERIC_1_L4, &sctp_mask,
-                              &sctp_val, sizeof(struct rte_sctp_hdr));
-       }
-
-       if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_UDP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_TCP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_SCTP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_OTHER) {
-               struct rte_ipv4_hdr ip4_mask, ip4_val;
-               memset(&ip4_mask, 0, sizeof(struct rte_ipv4_hdr));
-               memset(&ip4_val, 0, sizeof(struct rte_ipv4_hdr));
-
-               if (input->flow.ip4_flow.tos) {
-                       ip4_mask.type_of_service = masks->ipv4_mask.tos;
-                       ip4_val.type_of_service = input->flow.ip4_flow.tos;
-               }
-               if (input->flow.ip4_flow.ttl) {
-                       ip4_mask.time_to_live = masks->ipv4_mask.ttl;
-                       ip4_val.time_to_live = input->flow.ip4_flow.ttl;
-               }
-               if (input->flow.ip4_flow.proto) {
-                       ip4_mask.next_proto_id = masks->ipv4_mask.proto;
-                       ip4_val.next_proto_id = input->flow.ip4_flow.proto;
-               } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV4_SCTP) {
-                       /* Explicitly match the SCTP protocol number */
-                       ip4_mask.next_proto_id = 0xff;
-                       ip4_val.next_proto_id = IPPROTO_SCTP;
-               }
-               if (input->flow.ip4_flow.src_ip) {
-                       ip4_mask.src_addr =  masks->ipv4_mask.src_ip;
-                       ip4_val.src_addr = input->flow.ip4_flow.src_ip;
-               }
-               if (input->flow.ip4_flow.dst_ip) {
-                       ip4_mask.dst_addr =  masks->ipv4_mask.dst_ip;
-                       ip4_val.dst_addr = input->flow.ip4_flow.dst_ip;
-               }
-
-               enic_set_layer(gp, FILTER_GENERIC_1_IPV4, FILTER_GENERIC_1_L3,
-                       &ip4_mask, &ip4_val, sizeof(struct rte_ipv4_hdr));
-       }
-
-       if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_UDP) {
-               struct rte_udp_hdr udp_mask, udp_val;
-               memset(&udp_mask, 0, sizeof(udp_mask));
-               memset(&udp_val, 0, sizeof(udp_val));
-
-               if (input->flow.udp6_flow.src_port) {
-                       udp_mask.src_port = masks->src_port_mask;
-                       udp_val.src_port = input->flow.udp6_flow.src_port;
-               }
-               if (input->flow.udp6_flow.dst_port) {
-                       udp_mask.dst_port = masks->dst_port_mask;
-                       udp_val.dst_port = input->flow.udp6_flow.dst_port;
-               }
-               enic_set_layer(gp, FILTER_GENERIC_1_UDP, FILTER_GENERIC_1_L4,
-                              &udp_mask, &udp_val, sizeof(struct rte_udp_hdr));
-       } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_TCP) {
-               struct rte_tcp_hdr tcp_mask, tcp_val;
-               memset(&tcp_mask, 0, sizeof(tcp_mask));
-               memset(&tcp_val, 0, sizeof(tcp_val));
-
-               if (input->flow.tcp6_flow.src_port) {
-                       tcp_mask.src_port = masks->src_port_mask;
-                       tcp_val.src_port = input->flow.tcp6_flow.src_port;
-               }
-               if (input->flow.tcp6_flow.dst_port) {
-                       tcp_mask.dst_port = masks->dst_port_mask;
-                       tcp_val.dst_port = input->flow.tcp6_flow.dst_port;
-               }
-               enic_set_layer(gp, FILTER_GENERIC_1_TCP, FILTER_GENERIC_1_L4,
-                              &tcp_mask, &tcp_val, sizeof(struct rte_tcp_hdr));
-       } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_SCTP) {
-               struct rte_sctp_hdr sctp_mask, sctp_val;
-               memset(&sctp_mask, 0, sizeof(sctp_mask));
-               memset(&sctp_val, 0, sizeof(sctp_val));
-
-               if (input->flow.sctp6_flow.src_port) {
-                       sctp_mask.src_port = masks->src_port_mask;
-                       sctp_val.src_port = input->flow.sctp6_flow.src_port;
-               }
-               if (input->flow.sctp6_flow.dst_port) {
-                       sctp_mask.dst_port = masks->dst_port_mask;
-                       sctp_val.dst_port = input->flow.sctp6_flow.dst_port;
-               }
-               if (input->flow.sctp6_flow.verify_tag) {
-                       sctp_mask.tag = 0xffffffff;
-                       sctp_val.tag = input->flow.sctp6_flow.verify_tag;
-               }
-
-               enic_set_layer(gp, 0, FILTER_GENERIC_1_L4, &sctp_mask,
-                              &sctp_val, sizeof(struct rte_sctp_hdr));
-       }
-
-       if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_UDP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_TCP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_SCTP ||
-           input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_OTHER) {
-               struct rte_ipv6_hdr ipv6_mask, ipv6_val;
-               memset(&ipv6_mask, 0, sizeof(struct rte_ipv6_hdr));
-               memset(&ipv6_val, 0, sizeof(struct rte_ipv6_hdr));
-
-               if (input->flow.ipv6_flow.proto) {
-                       ipv6_mask.proto = masks->ipv6_mask.proto;
-                       ipv6_val.proto = input->flow.ipv6_flow.proto;
-               } else if (input->flow_type == RTE_ETH_FLOW_NONFRAG_IPV6_SCTP) {
-                       /* See comments for IPv4 SCTP above. */
-                       ipv6_mask.proto = 0xff;
-                       ipv6_val.proto = IPPROTO_SCTP;
-               }
-               memcpy(ipv6_mask.src_addr, masks->ipv6_mask.src_ip,
-                      sizeof(ipv6_mask.src_addr));
-               memcpy(ipv6_val.src_addr, input->flow.ipv6_flow.src_ip,
-                      sizeof(ipv6_val.src_addr));
-               memcpy(ipv6_mask.dst_addr, masks->ipv6_mask.dst_ip,
-                      sizeof(ipv6_mask.dst_addr));
-               memcpy(ipv6_val.dst_addr, input->flow.ipv6_flow.dst_ip,
-                      sizeof(ipv6_val.dst_addr));
-               if (input->flow.ipv6_flow.tc) {
-                       ipv6_mask.vtc_flow = masks->ipv6_mask.tc << 12;
-                       ipv6_val.vtc_flow = input->flow.ipv6_flow.tc << 12;
-               }
-               if (input->flow.ipv6_flow.hop_limits) {
-                       ipv6_mask.hop_limits = masks->ipv6_mask.hop_limits;
-                       ipv6_val.hop_limits = input->flow.ipv6_flow.hop_limits;
-               }
-
-               enic_set_layer(gp, FILTER_GENERIC_1_IPV6, FILTER_GENERIC_1_L3,
-                       &ipv6_mask, &ipv6_val, sizeof(struct rte_ipv6_hdr));
-       }
-}
-
-void enic_clsf_destroy(struct enic *enic)
-{
-       uint32_t index;
-       struct enic_fdir_node *key;
-       /* delete classifier entries */
-       for (index = 0; index < ENICPMD_FDIR_MAX; index++) {
-               key = enic->fdir.nodes[index];
-               if (key) {
-                       vnic_dev_classifier(enic->vdev, CLSF_DEL,
-                               &key->fltr_id, NULL, NULL);
-                       rte_free(key);
-                       enic->fdir.nodes[index] = NULL;
-               }
-       }
-
-       if (enic->fdir.hash) {
-               rte_hash_free(enic->fdir.hash);
-               enic->fdir.hash = NULL;
-       }
-}
-
-int enic_clsf_init(struct enic *enic)
-{
-       char clsf_name[RTE_HASH_NAMESIZE];
-       struct rte_hash_parameters hash_params = {
-               .name = clsf_name,
-               .entries = ENICPMD_CLSF_HASH_ENTRIES,
-               .key_len = sizeof(struct rte_eth_fdir_filter),
-               .hash_func = DEFAULT_HASH_FUNC,
-               .hash_func_init_val = 0,
-               .socket_id = SOCKET_ID_ANY,
-       };
-       snprintf(clsf_name, RTE_HASH_NAMESIZE, "enic_clsf_%s", enic->bdf_name);
-       enic->fdir.hash = rte_hash_create(&hash_params);
-       memset(&enic->fdir.stats, 0, sizeof(enic->fdir.stats));
-       enic->fdir.stats.free = ENICPMD_FDIR_MAX;
-       return NULL == enic->fdir.hash;
-}
index cebca7d..1eec96a 100644 (file)
@@ -1208,7 +1208,8 @@ enic_copy_action_v2(struct enic *enic,
                        const struct rte_flow_action_mark *mark =
                                (const struct rte_flow_action_mark *)
                                actions->conf;
-
+                       if (enic->use_noscatter_vec_rx_handler)
+                               return ENOTSUP;
                        if (overlap & MARK)
                                return ENOTSUP;
                        overlap |= MARK;
@@ -1228,6 +1229,8 @@ enic_copy_action_v2(struct enic *enic,
                        break;
                }
                case RTE_FLOW_ACTION_TYPE_FLAG: {
+                       if (enic->use_noscatter_vec_rx_handler)
+                               return ENOTSUP;
                        if (overlap & MARK)
                                return ENOTSUP;
                        overlap |= MARK;
index 86b91ed..ad1e89a 100644 (file)
@@ -1314,6 +1314,8 @@ enic_fm_copy_action(struct enic_flowman *fm,
                        const struct rte_flow_action_mark *mark =
                                actions->conf;
 
+                       if (enic->use_noscatter_vec_rx_handler)
+                               goto unsupported;
                        if (mark->id >= ENIC_MAGIC_FILTER_ID - 1)
                                return rte_flow_error_set(error, EINVAL,
                                        RTE_FLOW_ERROR_TYPE_ACTION,
@@ -1327,6 +1329,8 @@ enic_fm_copy_action(struct enic_flowman *fm,
                        break;
                }
                case RTE_FLOW_ACTION_TYPE_FLAG: {
+                       if (enic->use_noscatter_vec_rx_handler)
+                               goto unsupported;
                        /* ENIC_MAGIC_FILTER_ID is reserved for flagging */
                        memset(&fm_op, 0, sizeof(fm_op));
                        fm_op.fa_op = FMOP_MARK;
index d0d4103..755c0bf 100644 (file)
@@ -603,9 +603,6 @@ int enic_enable(struct enic *enic)
        err = enic_rxq_intr_init(enic);
        if (err)
                return err;
-       if (enic_clsf_init(enic))
-               dev_warning(enic, "Init of hash table for clsf failed."\
-                       "Flow director feature will not work\n");
 
        /* Initialize flowman if not already initialized during probe */
        if (enic->fm == NULL && enic_fm_init(enic))
@@ -1102,7 +1099,6 @@ int enic_disable(struct enic *enic)
 
        vnic_dev_disable(enic->vdev);
 
-       enic_clsf_destroy(enic);
        enic_fm_destroy(enic);
 
        if (!enic_is_sriov_vf(enic))
@@ -1753,9 +1749,6 @@ static int enic_dev_init(struct enic *enic)
                return -1;
        }
 
-       /* Get the supported filters */
-       enic_fdir_info(enic);
-
        eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr",
                                        sizeof(struct rte_ether_addr) *
                                        ENIC_UNICAST_PERFECT_FILTERS, 0);
index d39021f..f0bea33 100644 (file)
@@ -810,7 +810,6 @@ bool
 enic_use_vector_rx_handler(struct rte_eth_dev *eth_dev)
 {
        struct enic *enic = pmd_priv(eth_dev);
-       struct rte_fdir_conf *fconf;
 
        /* User needs to request for the avx2 handler */
        if (!enic->enable_avx2_rx)
@@ -818,10 +817,6 @@ enic_use_vector_rx_handler(struct rte_eth_dev *eth_dev)
        /* Do not support scatter Rx */
        if (!(enic->rq_count > 0 && enic->rq[0].data_queue_enable == 0))
                return false;
-       /* Do not support fdir/flow */
-       fconf = &eth_dev->data->dev_conf.fdir_conf;
-       if (fconf->mode != RTE_FDIR_MODE_NONE)
-               return false;
        if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_AVX2) &&
                        rte_vect_get_max_simd_bitwidth() >= RTE_VECT_SIMD_256) {
                ENICPMD_LOG(DEBUG, " use the non-scatter avx2 Rx handler");
index 86ef2a8..32c3312 100644 (file)
@@ -7,7 +7,6 @@ sources = files(
        'base/vnic_intr.c',
        'base/vnic_rq.c',
        'base/vnic_wq.c',
-       'enic_clsf.c',
        'enic_ethdev.c',
        'enic_flow.c',
        'enic_fm_flow.c',