net/ice: configure HW flow director rule
authorBeilei Xing <beilei.xing@intel.com>
Fri, 18 Oct 2019 11:15:55 +0000 (19:15 +0800)
committerFerruh Yigit <ferruh.yigit@intel.com>
Wed, 23 Oct 2019 14:43:10 +0000 (16:43 +0200)
This patch adds a HW FDIR rule to the FDIR HW table
without adding a FDIR filter.

Signed-off-by: Beilei Xing <beilei.xing@intel.com>
Acked-by: Qi Zhang <qi.z.zhang@intel.com>
Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>
drivers/net/ice/ice_ethdev.h
drivers/net/ice/ice_fdir_filter.c

index 222fc08..2060ad5 100644 (file)
@@ -281,6 +281,8 @@ struct ice_pf {
        uint16_t fdir_nb_qps; /* The number of queue pairs of Flow Director */
        uint16_t fdir_qp_offset;
        struct ice_fdir_info fdir; /* flow director info */
+       uint16_t hw_prof_cnt[ICE_FLTR_PTYPE_MAX][ICE_FD_HW_SEG_MAX];
+       uint16_t fdir_fltr_cnt[ICE_FLTR_PTYPE_MAX][ICE_FD_HW_SEG_MAX];
        struct ice_hw_port_stats stats_offset;
        struct ice_hw_port_stats stats;
        /* internal packet statistics, it should be excluded from the total */
index 0fb3054..60eb27e 100644 (file)
@@ -164,6 +164,56 @@ ice_fdir_prof_free(struct ice_hw *hw)
        rte_free(hw->fdir_prof);
 }
 
+/* Remove a profile for some filter type */
+static void
+ice_fdir_prof_rm(struct ice_pf *pf, enum ice_fltr_ptype ptype, bool is_tunnel)
+{
+       struct ice_hw *hw = ICE_PF_TO_HW(pf);
+       struct ice_fd_hw_prof *hw_prof;
+       uint64_t prof_id;
+       uint16_t vsi_num;
+       int i;
+
+       if (!hw->fdir_prof || !hw->fdir_prof[ptype])
+               return;
+
+       hw_prof = hw->fdir_prof[ptype];
+
+       prof_id = ptype + is_tunnel * ICE_FLTR_PTYPE_MAX;
+       for (i = 0; i < pf->hw_prof_cnt[ptype][is_tunnel]; i++) {
+               if (hw_prof->entry_h[i][is_tunnel]) {
+                       vsi_num = ice_get_hw_vsi_num(hw,
+                                                    hw_prof->vsi_h[i]);
+                       ice_rem_prof_id_flow(hw, ICE_BLK_FD,
+                                            vsi_num, ptype);
+                       ice_flow_rem_entry(hw,
+                                          hw_prof->entry_h[i][is_tunnel]);
+                       hw_prof->entry_h[i][is_tunnel] = 0;
+               }
+       }
+       ice_flow_rem_prof(hw, ICE_BLK_FD, prof_id);
+       rte_free(hw_prof->fdir_seg[is_tunnel]);
+       hw_prof->fdir_seg[is_tunnel] = NULL;
+
+       for (i = 0; i < hw_prof->cnt; i++)
+               hw_prof->vsi_h[i] = 0;
+       pf->hw_prof_cnt[ptype][is_tunnel] = 0;
+}
+
+/* Remove all created profiles */
+static void
+ice_fdir_prof_rm_all(struct ice_pf *pf)
+{
+       enum ice_fltr_ptype ptype;
+
+       for (ptype = ICE_FLTR_PTYPE_NONF_NONE;
+            ptype < ICE_FLTR_PTYPE_MAX;
+            ptype++) {
+               ice_fdir_prof_rm(pf, ptype, false);
+               ice_fdir_prof_rm(pf, ptype, true);
+       }
+}
+
 /*
  * ice_fdir_teardown - release the Flow Director resources
  * @pf: board private structure
@@ -192,9 +242,234 @@ ice_fdir_teardown(struct ice_pf *pf)
        pf->fdir.txq = NULL;
        ice_rx_queue_release(pf->fdir.rxq);
        pf->fdir.rxq = NULL;
+       ice_fdir_prof_rm_all(pf);
+       ice_fdir_prof_free(hw);
        ice_release_vsi(vsi);
        pf->fdir.fdir_vsi = NULL;
-       ice_fdir_prof_free(hw);
+}
+
+static int
+ice_fdir_hw_tbl_conf(struct ice_pf *pf, struct ice_vsi *vsi,
+                    struct ice_vsi *ctrl_vsi,
+                    struct ice_flow_seg_info *seg,
+                    enum ice_fltr_ptype ptype,
+                    bool is_tunnel)
+{
+       struct ice_hw *hw = ICE_PF_TO_HW(pf);
+       enum ice_flow_dir dir = ICE_FLOW_RX;
+       struct ice_flow_seg_info *ori_seg;
+       struct ice_fd_hw_prof *hw_prof;
+       struct ice_flow_prof *prof;
+       uint64_t entry_1 = 0;
+       uint64_t entry_2 = 0;
+       uint16_t vsi_num;
+       int ret;
+       uint64_t prof_id;
+
+       hw_prof = hw->fdir_prof[ptype];
+       ori_seg = hw_prof->fdir_seg[is_tunnel];
+       if (ori_seg) {
+               if (!is_tunnel) {
+                       if (!memcmp(ori_seg, seg, sizeof(*seg)))
+                               return -EAGAIN;
+               } else {
+                       if (!memcmp(ori_seg, &seg[1], sizeof(*seg)))
+                               return -EAGAIN;
+               }
+
+               if (pf->fdir_fltr_cnt[ptype][is_tunnel])
+                       return -EINVAL;
+
+               ice_fdir_prof_rm(pf, ptype, is_tunnel);
+       }
+
+       prof_id = ptype + is_tunnel * ICE_FLTR_PTYPE_MAX;
+       ret = ice_flow_add_prof(hw, ICE_BLK_FD, dir, prof_id, seg,
+                               (is_tunnel) ? 2 : 1, NULL, 0, &prof);
+       if (ret)
+               return ret;
+       ret = ice_flow_add_entry(hw, ICE_BLK_FD, prof_id, vsi->idx,
+                                vsi->idx, ICE_FLOW_PRIO_NORMAL,
+                                seg, NULL, 0, &entry_1);
+       if (ret) {
+               PMD_DRV_LOG(ERR, "Failed to add main VSI flow entry for %d.",
+                           ptype);
+               goto err_add_prof;
+       }
+       ret = ice_flow_add_entry(hw, ICE_BLK_FD, prof_id, vsi->idx,
+                                ctrl_vsi->idx, ICE_FLOW_PRIO_NORMAL,
+                                seg, NULL, 0, &entry_2);
+       if (ret) {
+               PMD_DRV_LOG(ERR, "Failed to add control VSI flow entry for %d.",
+                           ptype);
+               goto err_add_entry;
+       }
+
+       pf->hw_prof_cnt[ptype][is_tunnel] = 0;
+       hw_prof->cnt = 0;
+       hw_prof->fdir_seg[is_tunnel] = seg;
+       hw_prof->vsi_h[hw_prof->cnt] = vsi->idx;
+       hw_prof->entry_h[hw_prof->cnt++][is_tunnel] = entry_1;
+       pf->hw_prof_cnt[ptype][is_tunnel]++;
+       hw_prof->vsi_h[hw_prof->cnt] = ctrl_vsi->idx;
+       hw_prof->entry_h[hw_prof->cnt++][is_tunnel] = entry_2;
+       pf->hw_prof_cnt[ptype][is_tunnel]++;
+
+       return ret;
+
+err_add_entry:
+       vsi_num = ice_get_hw_vsi_num(hw, vsi->idx);
+       ice_rem_prof_id_flow(hw, ICE_BLK_FD, vsi_num, prof_id);
+       ice_flow_rem_entry(hw, entry_1);
+err_add_prof:
+       ice_flow_rem_prof(hw, ICE_BLK_FD, prof_id);
+
+       return ret;
+}
+
+static void
+ice_fdir_input_set_parse(uint64_t inset, enum ice_flow_field *field)
+{
+       uint32_t i, j;
+
+       struct ice_inset_map {
+               uint64_t inset;
+               enum ice_flow_field fld;
+       };
+       static const struct ice_inset_map ice_inset_map[] = {
+               {ICE_INSET_DMAC, ICE_FLOW_FIELD_IDX_ETH_DA},
+               {ICE_INSET_IPV4_SRC, ICE_FLOW_FIELD_IDX_IPV4_SA},
+               {ICE_INSET_IPV4_DST, ICE_FLOW_FIELD_IDX_IPV4_DA},
+               {ICE_INSET_IPV4_TOS, ICE_FLOW_FIELD_IDX_IPV4_DSCP},
+               {ICE_INSET_IPV4_TTL, ICE_FLOW_FIELD_IDX_IPV4_TTL},
+               {ICE_INSET_IPV4_PROTO, ICE_FLOW_FIELD_IDX_IPV4_PROT},
+               {ICE_INSET_IPV6_SRC, ICE_FLOW_FIELD_IDX_IPV6_SA},
+               {ICE_INSET_IPV6_DST, ICE_FLOW_FIELD_IDX_IPV6_DA},
+               {ICE_INSET_IPV6_TC, ICE_FLOW_FIELD_IDX_IPV6_DSCP},
+               {ICE_INSET_IPV6_NEXT_HDR, ICE_FLOW_FIELD_IDX_IPV6_PROT},
+               {ICE_INSET_IPV6_HOP_LIMIT, ICE_FLOW_FIELD_IDX_IPV6_TTL},
+               {ICE_INSET_TCP_SRC_PORT, ICE_FLOW_FIELD_IDX_TCP_SRC_PORT},
+               {ICE_INSET_TCP_DST_PORT, ICE_FLOW_FIELD_IDX_TCP_DST_PORT},
+               {ICE_INSET_UDP_SRC_PORT, ICE_FLOW_FIELD_IDX_UDP_SRC_PORT},
+               {ICE_INSET_UDP_DST_PORT, ICE_FLOW_FIELD_IDX_UDP_DST_PORT},
+               {ICE_INSET_SCTP_SRC_PORT, ICE_FLOW_FIELD_IDX_SCTP_SRC_PORT},
+               {ICE_INSET_SCTP_DST_PORT, ICE_FLOW_FIELD_IDX_SCTP_DST_PORT},
+       };
+
+       for (i = 0, j = 0; i < RTE_DIM(ice_inset_map); i++) {
+               if ((inset & ice_inset_map[i].inset) ==
+                   ice_inset_map[i].inset)
+                       field[j++] = ice_inset_map[i].fld;
+       }
+}
+
+static int __rte_unused
+ice_fdir_input_set_conf(struct ice_pf *pf, enum ice_fltr_ptype flow,
+                       uint64_t input_set, bool is_tunnel)
+{
+       struct ice_flow_seg_info *seg;
+       struct ice_flow_seg_info *seg_tun = NULL;
+       enum ice_flow_field field[ICE_FLOW_FIELD_IDX_MAX];
+       int i, ret;
+
+       if (!input_set)
+               return -EINVAL;
+
+       seg = (struct ice_flow_seg_info *)
+               ice_malloc(hw, sizeof(*seg));
+       if (!seg) {
+               PMD_DRV_LOG(ERR, "No memory can be allocated");
+               return -ENOMEM;
+       }
+
+       for (i = 0; i < ICE_FLOW_FIELD_IDX_MAX; i++)
+               field[i] = ICE_FLOW_FIELD_IDX_MAX;
+       ice_fdir_input_set_parse(input_set, field);
+
+       switch (flow) {
+       case ICE_FLTR_PTYPE_NONF_IPV4_UDP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_UDP |
+                                 ICE_FLOW_SEG_HDR_IPV4);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV4_TCP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_TCP |
+                                 ICE_FLOW_SEG_HDR_IPV4);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV4_SCTP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_SCTP |
+                                 ICE_FLOW_SEG_HDR_IPV4);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV4_OTHER:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_IPV4);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV6_UDP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_UDP |
+                                 ICE_FLOW_SEG_HDR_IPV6);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV6_TCP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_TCP |
+                                 ICE_FLOW_SEG_HDR_IPV6);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV6_SCTP:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_SCTP |
+                                 ICE_FLOW_SEG_HDR_IPV6);
+               break;
+       case ICE_FLTR_PTYPE_NONF_IPV6_OTHER:
+               ICE_FLOW_SET_HDRS(seg, ICE_FLOW_SEG_HDR_IPV6);
+               break;
+       default:
+               PMD_DRV_LOG(ERR, "not supported filter type.");
+               break;
+       }
+
+       for (i = 0; field[i] != ICE_FLOW_FIELD_IDX_MAX; i++) {
+               ice_flow_set_fld(seg, field[i],
+                                ICE_FLOW_FLD_OFF_INVAL,
+                                ICE_FLOW_FLD_OFF_INVAL,
+                                ICE_FLOW_FLD_OFF_INVAL, false);
+       }
+
+       if (!is_tunnel) {
+               ret = ice_fdir_hw_tbl_conf(pf, pf->main_vsi, pf->fdir.fdir_vsi,
+                                          seg, flow, false);
+       } else {
+               seg_tun = (struct ice_flow_seg_info *)
+                       ice_malloc(hw, sizeof(*seg) * ICE_FD_HW_SEG_MAX);
+               if (!seg_tun) {
+                       PMD_DRV_LOG(ERR, "No memory can be allocated");
+                       rte_free(seg);
+                       return -ENOMEM;
+               }
+               rte_memcpy(&seg_tun[1], seg, sizeof(*seg));
+               ret = ice_fdir_hw_tbl_conf(pf, pf->main_vsi, pf->fdir.fdir_vsi,
+                                          seg_tun, flow, true);
+       }
+
+       if (!ret) {
+               return ret;
+       } else if (ret < 0) {
+               rte_free(seg);
+               if (is_tunnel)
+                       rte_free(seg_tun);
+               return (ret == -EAGAIN) ? 0 : ret;
+       } else {
+               return ret;
+       }
+}
+
+static void __rte_unused
+ice_fdir_cnt_update(struct ice_pf *pf, enum ice_fltr_ptype ptype,
+                   bool is_tunnel, bool add)
+{
+       struct ice_hw *hw = ICE_PF_TO_HW(pf);
+       int cnt;
+
+       cnt = (add) ? 1 : -1;
+       hw->fdir_active_fltr += cnt;
+       if (ptype == ICE_FLTR_PTYPE_NONF_NONE || ptype >= ICE_FLTR_PTYPE_MAX)
+               PMD_DRV_LOG(ERR, "Unknown filter type %d", ptype);
+       else
+               pf->fdir_fltr_cnt[ptype][is_tunnel] += cnt;
 }
 
 static int