net/enic: fix hardcoding of some flow director masks
authorJohn Daley <johndale@cisco.com>
Thu, 9 Feb 2017 00:40:09 +0000 (16:40 -0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Fri, 10 Feb 2017 11:25:49 +0000 (12:25 +0100)
Hard coded mask values were being used for several of the IPv4 and IPv6
fields. Use the values in the rte_eth_fdir_masks structure provided by the
caller.

Fixes: dfbd6a9cb504 ("net/enic: extend flow director support for 1300 series")
Cc: stable@dpdk.org
Signed-off-by: John Daley <johndale@cisco.com>
drivers/net/enic/enic_clsf.c

index bcf479a..487f804 100644 (file)
@@ -211,15 +211,15 @@ copy_fltr_v2(struct filter_v2 *fltr, struct rte_eth_fdir_input *input,
                memset(&ip4_val, 0, sizeof(struct ipv4_hdr));
 
                if (input->flow.ip4_flow.tos) {
-                       ip4_mask.type_of_service = 0xff;
+                       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 = 0xff;
+                       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 = 0xff;
+                       ip4_mask.next_proto_id = masks->ipv4_mask.proto;
                        ip4_val.next_proto_id = input->flow.ip4_flow.proto;
                }
                if (input->flow.ip4_flow.src_ip) {
@@ -299,7 +299,7 @@ copy_fltr_v2(struct filter_v2 *fltr, struct rte_eth_fdir_input *input,
                memset(&ipv6_val, 0, sizeof(struct ipv6_hdr));
 
                if (input->flow.ipv6_flow.proto) {
-                       ipv6_mask.proto = 0xff;
+                       ipv6_mask.proto = masks->ipv6_mask.proto;
                        ipv6_val.proto = input->flow.ipv6_flow.proto;
                }
                for (i = 0; i < 4; i++) {
@@ -315,11 +315,11 @@ copy_fltr_v2(struct filter_v2 *fltr, struct rte_eth_fdir_input *input,
                                        input->flow.ipv6_flow.dst_ip[i];
                }
                if (input->flow.ipv6_flow.tc) {
-                       ipv6_mask.vtc_flow = 0x00ff0000;
-                       ipv6_val.vtc_flow = input->flow.ipv6_flow.tc << 16;
+                       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 = 0xff;
+                       ipv6_mask.hop_limits = masks->ipv6_mask.hop_limits;
                        ipv6_val.hop_limits = input->flow.ipv6_flow.hop_limits;
                }