net/enic: fix bad L4 checksum flag on ICMP packets
authorJohn Daley <johndale@cisco.com>
Wed, 17 Aug 2016 22:15:26 +0000 (15:15 -0700)
committerBruce Richardson <bruce.richardson@intel.com>
Fri, 30 Sep 2016 10:27:18 +0000 (12:27 +0200)
The bad L4 checksum flag was set on IP packets which were not
also TCP or UDP packets. This includes ICMP, IGMP and OSPF packets.

L4 ptypes were being treated as bits instead of values within the
L4 mask causing the code to check L4 checksum in the completion
queue and incorrectly set the L4 bad checksum flag.

Fixes: 947d860c821f ("enic: improve Rx performance")

Reviewed-by: Nelson Escobar <neescoba@cisco.com>
Signed-off-by: John Daley <johndale@cisco.com>
drivers/net/enic/enic_rxtx.c

index 50f0b28..ad59613 100644 (file)
@@ -212,9 +212,12 @@ enic_cq_rx_to_pkt_flags(struct cq_desc *cqd, struct rte_mbuf *mbuf)
        /* checksum flags */
        if (!enic_cq_rx_desc_csum_not_calc(cqrd) &&
                (mbuf->packet_type & RTE_PTYPE_L3_IPV4)) {
+               uint32_t l4_flags = mbuf->packet_type & RTE_PTYPE_L4_MASK;
+
                if (unlikely(!enic_cq_rx_desc_ipv4_csum_ok(cqrd)))
                        pkt_flags |= PKT_RX_IP_CKSUM_BAD;
-               if (mbuf->packet_type & (RTE_PTYPE_L4_UDP | RTE_PTYPE_L4_TCP)) {
+               if (l4_flags == RTE_PTYPE_L4_UDP ||
+                   l4_flags == RTE_PTYPE_L4_TCP) {
                        if (unlikely(!enic_cq_rx_desc_tcp_udp_csum_ok(cqrd)))
                                pkt_flags |= PKT_RX_L4_CKSUM_BAD;
                }