app/testpmd: configure flow director flexible mask
authorJingjing Wu <jingjing.wu@intel.com>
Fri, 21 Nov 2014 00:46:55 +0000 (08:46 +0800)
committerThomas Monjalon <thomas.monjalon@6wind.com>
Mon, 24 Nov 2014 23:06:39 +0000 (00:06 +0100)
test command added to configure flexible mask

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
app/test-pmd/cmdline.c
app/test-pmd/config.c
app/test-pmd/testpmd.h

index b5d65c3..f527b63 100644 (file)
@@ -723,6 +723,11 @@ static void cmd_help_long_parsed(void *parsed_result,
 
                        "flush_flow_director (port_id)\n"
                        "    Flush all flow director entries of a device.\n\n"
+
+                       "flow_director_flex_mask (port_id)"
+                       " flow (ip4|ip4-frag|tcp4|udp4|sctp4|ip6|ip6-frag|tcp6|udp6|sctp6|all)"
+                       " (mask)\n"
+                       "    Configure mask of flex payload.\n\n"
                );
        }
 }
@@ -8312,6 +8317,94 @@ cmdline_parse_inst_t cmd_flush_flow_director = {
        },
 };
 
+/* *** deal with flow director mask on flexible payload *** */
+struct cmd_flow_director_flex_mask_result {
+       cmdline_fixed_string_t flow_director_flexmask;
+       uint8_t port_id;
+       cmdline_fixed_string_t flow;
+       cmdline_fixed_string_t flow_type;
+       cmdline_fixed_string_t mask;
+};
+
+static void
+cmd_flow_director_flex_mask_parsed(void *parsed_result,
+                         __attribute__((unused)) struct cmdline *cl,
+                         __attribute__((unused)) void *data)
+{
+       struct cmd_flow_director_flex_mask_result *res = parsed_result;
+       struct rte_eth_fdir_flex_mask flex_mask;
+       struct rte_port *port;
+       enum rte_eth_flow_type i;
+       int ret;
+
+       if (res->port_id > nb_ports) {
+               printf("Invalid port, range is [0, %d]\n", nb_ports - 1);
+               return;
+       }
+
+       port = &ports[res->port_id];
+       /** Check if the port is not started **/
+       if (port->port_status != RTE_PORT_STOPPED) {
+               printf("Please stop port %d first\n", res->port_id);
+               return;
+       }
+
+       memset(&flex_mask, 0, sizeof(struct rte_eth_fdir_flex_mask));
+       ret = parse_flexbytes(res->mask,
+                       flex_mask.mask,
+                       RTE_ETH_FDIR_MAX_FLEXLEN);
+       if (ret < 0) {
+               printf("error: Cannot parse mask input.\n");
+               return;
+       }
+       if (!strcmp(res->flow_type, "all")) {
+               for (i = RTE_ETH_FLOW_TYPE_UDPV4;
+                    i <= RTE_ETH_FLOW_TYPE_FRAG_IPV6;
+                    i++) {
+                       flex_mask.flow_type = i;
+                       fdir_set_flex_mask(res->port_id, &flex_mask);
+               }
+               cmd_reconfig_device_queue(res->port_id, 1, 0);
+               return;
+       }
+       flex_mask.flow_type = str2flowtype(res->flow_type);
+       fdir_set_flex_mask(res->port_id, &flex_mask);
+       cmd_reconfig_device_queue(res->port_id, 1, 0);
+}
+
+cmdline_parse_token_string_t cmd_flow_director_flexmask =
+       TOKEN_STRING_INITIALIZER(struct cmd_flow_director_flex_mask_result,
+                                flow_director_flexmask,
+                                "flow_director_flex_mask");
+cmdline_parse_token_num_t cmd_flow_director_flexmask_port_id =
+       TOKEN_NUM_INITIALIZER(struct cmd_flow_director_flex_mask_result,
+                             port_id, UINT8);
+cmdline_parse_token_string_t cmd_flow_director_flexmask_flow =
+       TOKEN_STRING_INITIALIZER(struct cmd_flow_director_flex_mask_result,
+                                flow, "flow");
+cmdline_parse_token_string_t cmd_flow_director_flexmask_flow_type =
+       TOKEN_STRING_INITIALIZER(struct cmd_flow_director_flex_mask_result,
+                                flow_type,
+                               "ip4#ip4-frag#tcp4#udp4#sctp4#"
+                               "ip6#ip6-frag#tcp6#udp6#sctp6#all");
+cmdline_parse_token_string_t cmd_flow_director_flexmask_mask =
+       TOKEN_STRING_INITIALIZER(struct cmd_flow_director_flex_mask_result,
+                                mask, NULL);
+
+cmdline_parse_inst_t cmd_set_flow_director_flex_mask = {
+       .f = cmd_flow_director_flex_mask_parsed,
+       .data = NULL,
+       .help_str = "set flow director's flex mask on NIC",
+       .tokens = {
+               (void *)&cmd_flow_director_flexmask,
+               (void *)&cmd_flow_director_flexmask_port_id,
+               (void *)&cmd_flow_director_flexmask_flow,
+               (void *)&cmd_flow_director_flexmask_flow_type,
+               (void *)&cmd_flow_director_flexmask_mask,
+               NULL,
+       },
+};
+
 /* ******************************************************************************** */
 
 /* list of instructions */
@@ -8446,6 +8539,7 @@ cmdline_parse_ctx_t main_ctx[] = {
        (cmdline_parse_inst_t *)&cmd_add_del_udp_flow_director,
        (cmdline_parse_inst_t *)&cmd_add_del_sctp_flow_director,
        (cmdline_parse_inst_t *)&cmd_flush_flow_director,
+       (cmdline_parse_inst_t *)&cmd_set_flow_director_flex_mask,
        NULL,
 };
 
index 1dda5e4..eb26306 100644 (file)
@@ -2019,6 +2019,36 @@ fdir_set_masks(portid_t port_id, struct rte_fdir_masks *fdir_masks)
               "diag=%d\n", port_id, diag);
 }
 
+void
+fdir_set_flex_mask(portid_t port_id, struct rte_eth_fdir_flex_mask *cfg)
+{
+       struct rte_port *port;
+       struct rte_eth_fdir_flex_conf *flex_conf;
+       int i, idx = 0;
+
+       port = &ports[port_id];
+       flex_conf = &port->dev_conf.fdir_conf.flex_conf;
+       for (i = 0; i < RTE_ETH_FLOW_TYPE_MAX; i++) {
+               if (cfg->flow_type == flex_conf->flex_mask[i].flow_type) {
+                       idx = i;
+                       break;
+               }
+       }
+       if (i >= RTE_ETH_FLOW_TYPE_MAX) {
+               if (flex_conf->nb_flexmasks < RTE_DIM(flex_conf->flex_mask)) {
+                       idx = flex_conf->nb_flexmasks;
+                       flex_conf->nb_flexmasks++;
+               } else {
+                       printf("The flex mask table is full. Can not set flex"
+                               " mask for flow_type(%u).", cfg->flow_type);
+                       return;
+               }
+       }
+       (void)rte_memcpy(&flex_conf->flex_mask[idx],
+                        cfg,
+                        sizeof(struct rte_eth_fdir_flex_mask));
+}
+
 void
 set_vf_traffic(portid_t port_id, uint8_t is_rx, uint16_t vf, uint8_t on)
 {
index 10bf13f..b390874 100644 (file)
@@ -532,7 +532,8 @@ void fdir_update_perfect_filter(portid_t port_id, uint16_t soft_id,
 void fdir_remove_perfect_filter(portid_t port_id, uint16_t soft_id,
                                struct rte_fdir_filter *fdir_filter);
 void fdir_set_masks(portid_t port_id, struct rte_fdir_masks *fdir_masks);
-
+void fdir_set_flex_mask(portid_t port_id,
+                          struct rte_eth_fdir_flex_mask *cfg);
 void port_rss_reta_info(portid_t port_id,
                        struct rte_eth_rss_reta_entry64 *reta_conf,
                        uint16_t nb_entries);