net/enic: use TSO flags
authorJohn Daley <johndale@cisco.com>
Wed, 10 Jan 2018 09:17:11 +0000 (01:17 -0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Tue, 16 Jan 2018 17:47:49 +0000 (18:47 +0100)
Depend on the tx_offload flags in the mbuf to determine the length
of the headers instead of looking into the packet itself.

Signed-off-by: John Daley <johndale@cisco.com>
Reviewed-by: Hyong Youb Kim <hyonkim@cisco.com>
drivers/net/enic/enic_rxtx.c

index b9358d3..08d600f 100644 (file)
@@ -105,59 +105,6 @@ enic_cq_rx_desc_n_bytes(struct cq_desc *cqd)
                CQ_ENET_RQ_DESC_BYTES_WRITTEN_MASK;
 }
 
-/* Find the offset to L5. This is needed by enic TSO implementation.
- * Return 0 if not a TCP packet or can't figure out the length.
- */
-static inline uint8_t tso_header_len(struct rte_mbuf *mbuf)
-{
-       struct ether_hdr *eh;
-       struct vlan_hdr *vh;
-       struct ipv4_hdr *ip4;
-       struct ipv6_hdr *ip6;
-       struct tcp_hdr *th;
-       uint8_t hdr_len;
-       uint16_t ether_type;
-
-       /* offset past Ethernet header */
-       eh = rte_pktmbuf_mtod(mbuf, struct ether_hdr *);
-       ether_type = eh->ether_type;
-       hdr_len = sizeof(struct ether_hdr);
-       if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_VLAN)) {
-               vh = rte_pktmbuf_mtod_offset(mbuf, struct vlan_hdr *, hdr_len);
-               ether_type = vh->eth_proto;
-               hdr_len += sizeof(struct vlan_hdr);
-       }
-
-       /* offset past IP header */
-       switch (rte_be_to_cpu_16(ether_type)) {
-       case ETHER_TYPE_IPv4:
-               ip4 = rte_pktmbuf_mtod_offset(mbuf, struct ipv4_hdr *, hdr_len);
-               if (ip4->next_proto_id != IPPROTO_TCP)
-                       return 0;
-               hdr_len += (ip4->version_ihl & 0xf) * 4;
-               break;
-       case ETHER_TYPE_IPv6:
-               ip6 = rte_pktmbuf_mtod_offset(mbuf, struct ipv6_hdr *, hdr_len);
-               if (ip6->proto != IPPROTO_TCP)
-                       return 0;
-               hdr_len += sizeof(struct ipv6_hdr);
-               break;
-       default:
-               return 0;
-       }
-
-       if ((hdr_len + sizeof(struct tcp_hdr)) > mbuf->pkt_len)
-               return 0;
-
-       /* offset past TCP header */
-       th = rte_pktmbuf_mtod_offset(mbuf, struct tcp_hdr *, hdr_len);
-       hdr_len += (th->data_off >> 4) * 4;
-
-       if (hdr_len > mbuf->pkt_len)
-               return 0;
-
-       return hdr_len;
-}
 
 static inline uint8_t
 enic_cq_rx_check_err(struct cq_desc *cqd)
@@ -556,7 +503,8 @@ uint16_t enic_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
                header_len = 0;
 
                if (tso) {
-                       header_len = tso_header_len(tx_pkt);
+                       header_len = tx_pkt->l2_len + tx_pkt->l3_len +
+                                    tx_pkt->l4_len;
 
                        /* Drop if non-TCP packet or TSO seg size is too big */
                        if (unlikely(header_len == 0 || ((tx_pkt->tso_segsz +