2 * Copyright (c) 2015, Mellanox Technologies. All rights reserved.
4 * This software is available to you under a choice of one of two
5 * licenses. You may choose to be licensed under the terms of the GNU
6 * General Public License (GPL) Version 2, available from the file
7 * COPYING in the main directory of this source tree, or the
8 * OpenIB.org BSD license below:
10 * Redistribution and use in source and binary forms, with or
11 * without modification, are permitted provided that the following
14 * - Redistributions of source code must retain the above
15 * copyright notice, this list of conditions and the following
18 * - Redistributions in binary form must reproduce the above
19 * copyright notice, this list of conditions and the following
20 * disclaimer in the documentation and/or other materials
21 * provided with the distribution.
23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
24 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
25 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
26 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
27 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
28 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
29 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
33 #ifndef _MLX5_FS_CORE_
34 #define _MLX5_FS_CORE_
36 #include <linux/refcount.h>
37 #include <linux/mlx5/fs.h>
38 #include <linux/rhashtable.h>
39 #include <linux/llist.h>
41 /* FS_TYPE_PRIO_CHAINS is a PRIO that will have namespaces only,
42 * and those are in parallel to one another when going over them to connect
43 * a new flow table. Meaning the last flow table in a TYPE_PRIO prio in one
44 * parallel namespace will not automatically connect to the first flow table
45 * found in any prio in any next namespace, but skip the entire containing
46 * TYPE_PRIO_CHAINS prio.
48 * This is used to implement tc chains, each chain of prios is a different
49 * namespace inside a containing TYPE_PRIO_CHAINS prio.
62 enum fs_flow_table_type {
65 FS_FT_ESW_EGRESS_ACL = 0x2,
66 FS_FT_ESW_INGRESS_ACL = 0x3,
68 FS_FT_SNIFFER_RX = 0X5,
69 FS_FT_SNIFFER_TX = 0X6,
71 FS_FT_MAX_TYPE = FS_FT_RDMA_RX,
74 enum fs_flow_table_op_mod {
76 FS_FT_OP_MOD_LAG_DEMUX,
80 FS_FTE_STATUS_EXISTING = 1UL << 0,
83 struct mlx5_flow_steering {
84 struct mlx5_core_dev *dev;
85 struct kmem_cache *fgs_cache;
86 struct kmem_cache *ftes_cache;
87 struct mlx5_flow_root_namespace *root_ns;
88 struct mlx5_flow_root_namespace *fdb_root_ns;
89 struct mlx5_flow_namespace **fdb_sub_ns;
90 struct mlx5_flow_root_namespace **esw_egress_root_ns;
91 struct mlx5_flow_root_namespace **esw_ingress_root_ns;
92 struct mlx5_flow_root_namespace *sniffer_tx_root_ns;
93 struct mlx5_flow_root_namespace *sniffer_rx_root_ns;
94 struct mlx5_flow_root_namespace *rdma_rx_root_ns;
95 struct mlx5_flow_root_namespace *egress_root_ns;
99 struct list_head list;
100 struct list_head children;
101 enum fs_node_type type;
102 struct fs_node *parent;
103 struct fs_node *root;
104 /* lock the node for writing and traversing */
105 struct rw_semaphore lock;
108 void (*del_hw_func)(struct fs_node *);
109 void (*del_sw_func)(struct fs_node *);
113 struct mlx5_flow_rule {
115 struct mlx5_flow_destination dest_attr;
116 /* next_ft should be accessed under chain_lock and only of
117 * destination type is FWD_NEXT_fT.
119 struct list_head next_ft;
123 struct mlx5_flow_handle {
125 struct mlx5_flow_rule *rule[];
128 /* Type of children is mlx5_flow_group */
129 struct mlx5_flow_table {
133 unsigned int max_fte;
135 enum fs_flow_table_type type;
136 enum fs_flow_table_op_mod op_mod;
139 unsigned int required_groups;
140 unsigned int num_groups;
142 /* Protect fwd_rules */
144 /* FWD rules that point on this flow table */
145 struct list_head fwd_rules;
147 struct rhltable fgs_hash;
148 enum mlx5_flow_table_miss_action def_miss_action;
151 struct mlx5_ft_underlay_qp {
152 struct list_head list;
156 #define MLX5_FTE_MATCH_PARAM_RESERVED reserved_at_a00
157 /* Calculate the fte_match_param length and without the reserved length.
158 * Make sure the reserved field is the last.
160 #define MLX5_ST_SZ_DW_MATCH_PARAM \
161 ((MLX5_BYTE_OFF(fte_match_param, MLX5_FTE_MATCH_PARAM_RESERVED) / sizeof(u32)) + \
162 BUILD_BUG_ON_ZERO(MLX5_ST_SZ_BYTES(fte_match_param) != \
163 MLX5_FLD_SZ_BYTES(fte_match_param, \
164 MLX5_FTE_MATCH_PARAM_RESERVED) +\
165 MLX5_BYTE_OFF(fte_match_param, \
166 MLX5_FTE_MATCH_PARAM_RESERVED)))
168 /* Type of children is mlx5_flow_rule */
171 u32 val[MLX5_ST_SZ_DW_MATCH_PARAM];
174 struct mlx5_flow_context flow_context;
175 struct mlx5_flow_act action;
176 enum fs_fte_status status;
177 struct mlx5_fc *counter;
178 struct rhash_head hash;
182 /* Type of children is mlx5_flow_table/namespace */
185 unsigned int num_levels;
186 unsigned int start_level;
191 /* Type of children is fs_prio */
192 struct mlx5_flow_namespace {
193 /* parent == NULL => root ns */
195 enum mlx5_flow_table_miss_action def_miss_action;
198 struct mlx5_flow_group_mask {
199 u8 match_criteria_enable;
200 u32 match_criteria[MLX5_ST_SZ_DW_MATCH_PARAM];
203 /* Type of children is fs_fte */
204 struct mlx5_flow_group {
206 struct mlx5_flow_group_mask mask;
209 struct ida fte_allocator;
211 struct rhashtable ftes_hash;
212 struct rhlist_head hash;
215 struct mlx5_flow_root_namespace {
216 struct mlx5_flow_namespace ns;
217 enum fs_flow_table_type table_type;
218 struct mlx5_core_dev *dev;
219 struct mlx5_flow_table *root_ft;
220 /* Should be held when chaining flow tables */
221 struct mutex chain_lock;
222 struct list_head underlay_qpns;
223 const struct mlx5_flow_cmds *cmds;
226 int mlx5_init_fc_stats(struct mlx5_core_dev *dev);
227 void mlx5_cleanup_fc_stats(struct mlx5_core_dev *dev);
228 void mlx5_fc_queue_stats_work(struct mlx5_core_dev *dev,
229 struct delayed_work *dwork,
230 unsigned long delay);
231 void mlx5_fc_update_sampling_interval(struct mlx5_core_dev *dev,
232 unsigned long interval);
234 int mlx5_init_fs(struct mlx5_core_dev *dev);
235 void mlx5_cleanup_fs(struct mlx5_core_dev *dev);
237 #define fs_get_obj(v, _node) {v = container_of((_node), typeof(*v), node); }
239 #define fs_list_for_each_entry(pos, root) \
240 list_for_each_entry(pos, root, node.list)
242 #define fs_list_for_each_entry_safe(pos, tmp, root) \
243 list_for_each_entry_safe(pos, tmp, root, node.list)
245 #define fs_for_each_ns_or_ft_reverse(pos, prio) \
246 list_for_each_entry_reverse(pos, &(prio)->node.children, list)
248 #define fs_for_each_ns_or_ft(pos, prio) \
249 list_for_each_entry(pos, (&(prio)->node.children), list)
251 #define fs_for_each_prio(pos, ns) \
252 fs_list_for_each_entry(pos, &(ns)->node.children)
254 #define fs_for_each_ns(pos, prio) \
255 fs_list_for_each_entry(pos, &(prio)->node.children)
257 #define fs_for_each_ft(pos, prio) \
258 fs_list_for_each_entry(pos, &(prio)->node.children)
260 #define fs_for_each_ft_safe(pos, tmp, prio) \
261 fs_list_for_each_entry_safe(pos, tmp, &(prio)->node.children)
263 #define fs_for_each_fg(pos, ft) \
264 fs_list_for_each_entry(pos, &(ft)->node.children)
266 #define fs_for_each_fte(pos, fg) \
267 fs_list_for_each_entry(pos, &(fg)->node.children)
269 #define fs_for_each_dst(pos, fte) \
270 fs_list_for_each_entry(pos, &(fte)->node.children)
272 #define MLX5_CAP_FLOWTABLE_TYPE(mdev, cap, type) ( \
273 (type == FS_FT_NIC_RX) ? MLX5_CAP_FLOWTABLE_NIC_RX(mdev, cap) : \
274 (type == FS_FT_ESW_EGRESS_ACL) ? MLX5_CAP_ESW_EGRESS_ACL(mdev, cap) : \
275 (type == FS_FT_ESW_INGRESS_ACL) ? MLX5_CAP_ESW_INGRESS_ACL(mdev, cap) : \
276 (type == FS_FT_FDB) ? MLX5_CAP_ESW_FLOWTABLE_FDB(mdev, cap) : \
277 (type == FS_FT_SNIFFER_RX) ? MLX5_CAP_FLOWTABLE_SNIFFER_RX(mdev, cap) : \
278 (type == FS_FT_SNIFFER_TX) ? MLX5_CAP_FLOWTABLE_SNIFFER_TX(mdev, cap) : \
279 (type == FS_FT_RDMA_RX) ? MLX5_CAP_FLOWTABLE_RDMA_RX(mdev, cap) : \
280 (BUILD_BUG_ON_ZERO(FS_FT_RDMA_RX != FS_FT_MAX_TYPE))\