net/ixgbe: enable IPv6 for consistent API
authorQi Zhang <qi.z.zhang@intel.com>
Wed, 5 Jul 2017 02:21:36 +0000 (22:21 -0400)
committerFerruh Yigit <ferruh.yigit@intel.com>
Thu, 6 Jul 2017 13:00:57 +0000 (15:00 +0200)
Enable IPv6 support with rte_flow API.
Only supports signature Match.

Signed-off-by: Qi Zhang <qi.z.zhang@intel.com>
Acked-by: Wenzhuo Lu <wenzhuo.lu@intel.com>
drivers/net/ixgbe/ixgbe_flow.c

index 593f267..6885abd 100644 (file)
@@ -1333,12 +1333,13 @@ static inline uint8_t signature_match(const struct rte_flow_item pattern[])
  * Parse the rule to see if it is a IP or MAC VLAN flow director rule.
  * And get the flow director filter info BTW.
  * UDP/TCP/SCTP PATTERN:
- * The first not void item can be ETH or IPV4.
- * The second not void item must be IPV4 if the first one is ETH.
+ * The first not void item can be ETH or IPV4 or IPV6
+ * The second not void item must be IPV4 or IPV6 if the first one is ETH.
  * The next not void item could be UDP or TCP or SCTP (optional)
  * The next not void item could be RAW (for flexbyte, optional)
  * The next not void item must be END.
- * A Fuzzy Match pattern can appear at any place before END (optional)
+ * A Fuzzy Match pattern can appear at any place before END.
+ * Fuzzy Match is optional for IPV4 but is required for IPV6
  * MAC VLAN PATTERN:
  * The first not void item must be ETH.
  * The second not void item must be MAC VLAN.
@@ -1386,6 +1387,8 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
        const struct rte_flow_item_eth *eth_mask;
        const struct rte_flow_item_ipv4 *ipv4_spec;
        const struct rte_flow_item_ipv4 *ipv4_mask;
+       const struct rte_flow_item_ipv6 *ipv6_spec;
+       const struct rte_flow_item_ipv6 *ipv6_mask;
        const struct rte_flow_item_tcp *tcp_spec;
        const struct rte_flow_item_tcp *tcp_mask;
        const struct rte_flow_item_udp *udp_spec;
@@ -1397,7 +1400,7 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
        const struct rte_flow_item_raw *raw_mask;
        const struct rte_flow_item_raw *raw_spec;
 
-       uint32_t j;
+       uint8_t j;
 
        if (!pattern) {
                rte_flow_error_set(error, EINVAL,
@@ -1436,6 +1439,7 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
        item = next_no_fuzzy_pattern(pattern, NULL);
        if (item->type != RTE_FLOW_ITEM_TYPE_ETH &&
            item->type != RTE_FLOW_ITEM_TYPE_IPV4 &&
+           item->type != RTE_FLOW_ITEM_TYPE_IPV6 &&
            item->type != RTE_FLOW_ITEM_TYPE_TCP &&
            item->type != RTE_FLOW_ITEM_TYPE_UDP &&
            item->type != RTE_FLOW_ITEM_TYPE_SCTP) {
@@ -1588,7 +1592,7 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
                }
        }
 
-       /* Get the IP info. */
+       /* Get the IPV4 info. */
        if (item->type == RTE_FLOW_ITEM_TYPE_IPV4) {
                /**
                 * Set the flow type even if there's no content
@@ -1662,14 +1666,106 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
                }
        }
 
+       /* Get the IPV6 info. */
+       if (item->type == RTE_FLOW_ITEM_TYPE_IPV6) {
+               /**
+                * Set the flow type even if there's no content
+                * as we must have a flow type.
+                */
+               rule->ixgbe_fdir.formatted.flow_type =
+                       IXGBE_ATR_FLOW_TYPE_IPV6;
+
+               /**
+                * 1. must signature match
+                * 2. not support last
+                * 3. mask must not null
+                */
+               if (rule->mode != RTE_FDIR_MODE_SIGNATURE ||
+                   item->last ||
+                   !item->mask) {
+                       memset(rule, 0, sizeof(struct ixgbe_fdir_rule));
+                       rte_flow_error_set(error, EINVAL,
+                               RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+                               item, "Not supported last point for range");
+                       return -rte_errno;
+               }
+
+               rule->b_mask = TRUE;
+               ipv6_mask =
+                       (const struct rte_flow_item_ipv6 *)item->mask;
+               if (ipv6_mask->hdr.vtc_flow ||
+                   ipv6_mask->hdr.payload_len ||
+                   ipv6_mask->hdr.proto ||
+                   ipv6_mask->hdr.hop_limits) {
+                       memset(rule, 0, sizeof(struct ixgbe_fdir_rule));
+                       rte_flow_error_set(error, EINVAL,
+                               RTE_FLOW_ERROR_TYPE_ITEM,
+                               item, "Not supported by fdir filter");
+                       return -rte_errno;
+               }
+
+               /* check src addr mask */
+               for (j = 0; j < 16; j++) {
+                       if (ipv6_mask->hdr.src_addr[j] == UINT8_MAX) {
+                               rule->mask.src_ipv6_mask |= 1 << j;
+                       } else if (ipv6_mask->hdr.src_addr[j] != 0) {
+                               memset(rule, 0, sizeof(struct ixgbe_fdir_rule));
+                               rte_flow_error_set(error, EINVAL,
+                                       RTE_FLOW_ERROR_TYPE_ITEM,
+                                       item, "Not supported by fdir filter");
+                               return -rte_errno;
+                       }
+               }
+
+               /* check dst addr mask */
+               for (j = 0; j < 16; j++) {
+                       if (ipv6_mask->hdr.dst_addr[j] == UINT8_MAX) {
+                               rule->mask.dst_ipv6_mask |= 1 << j;
+                       } else if (ipv6_mask->hdr.dst_addr[j] != 0) {
+                               memset(rule, 0, sizeof(struct ixgbe_fdir_rule));
+                               rte_flow_error_set(error, EINVAL,
+                                       RTE_FLOW_ERROR_TYPE_ITEM,
+                                       item, "Not supported by fdir filter");
+                               return -rte_errno;
+                       }
+               }
+
+               if (item->spec) {
+                       rule->b_spec = TRUE;
+                       ipv6_spec =
+                               (const struct rte_flow_item_ipv6 *)item->spec;
+                       rte_memcpy(rule->ixgbe_fdir.formatted.src_ip,
+                                  ipv6_spec->hdr.src_addr, 16);
+                       rte_memcpy(rule->ixgbe_fdir.formatted.dst_ip,
+                                  ipv6_spec->hdr.dst_addr, 16);
+               }
+
+               /**
+                * Check if the next not void item is
+                * TCP or UDP or SCTP or END.
+                */
+               item = next_no_fuzzy_pattern(pattern, item);
+               if (item->type != RTE_FLOW_ITEM_TYPE_TCP &&
+                   item->type != RTE_FLOW_ITEM_TYPE_UDP &&
+                   item->type != RTE_FLOW_ITEM_TYPE_SCTP &&
+                   item->type != RTE_FLOW_ITEM_TYPE_END &&
+                   item->type != RTE_FLOW_ITEM_TYPE_RAW) {
+                       memset(rule, 0, sizeof(struct ixgbe_fdir_rule));
+                       rte_flow_error_set(error, EINVAL,
+                               RTE_FLOW_ERROR_TYPE_ITEM,
+                               item, "Not supported by fdir filter");
+                       return -rte_errno;
+               }
+       }
+
        /* Get the TCP info. */
        if (item->type == RTE_FLOW_ITEM_TYPE_TCP) {
                /**
                 * Set the flow type even if there's no content
                 * as we must have a flow type.
                 */
-               rule->ixgbe_fdir.formatted.flow_type =
-                       IXGBE_ATR_FLOW_TYPE_TCPV4;
+               rule->ixgbe_fdir.formatted.flow_type |=
+                       IXGBE_ATR_L4TYPE_TCP;
                /*Not supported last point for range*/
                if (item->last) {
                        rte_flow_error_set(error, EINVAL,
@@ -1733,8 +1829,8 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
                 * Set the flow type even if there's no content
                 * as we must have a flow type.
                 */
-               rule->ixgbe_fdir.formatted.flow_type =
-                       IXGBE_ATR_FLOW_TYPE_UDPV4;
+               rule->ixgbe_fdir.formatted.flow_type |=
+                       IXGBE_ATR_L4TYPE_UDP;
                /*Not supported last point for range*/
                if (item->last) {
                        rte_flow_error_set(error, EINVAL,
@@ -1793,8 +1889,8 @@ ixgbe_parse_fdir_filter_normal(const struct rte_flow_attr *attr,
                 * Set the flow type even if there's no content
                 * as we must have a flow type.
                 */
-               rule->ixgbe_fdir.formatted.flow_type =
-                       IXGBE_ATR_FLOW_TYPE_SCTPV4;
+               rule->ixgbe_fdir.formatted.flow_type |=
+                       IXGBE_ATR_L4TYPE_SCTP;
                /*Not supported last point for range*/
                if (item->last) {
                        rte_flow_error_set(error, EINVAL,